StreamPipesClientConfig(credential_provider, host_address, https_disabled=False, port=80)
+ StreamPipesClientConfig(credential_provider, host_address, https_disabled=False, port=80, additional_headers=dict())
dataclass
@@ -2381,6 +2381,24 @@ Specifies additional HTTP headers that should be sent with each request, e.g., proxy headers
+
+
+ TYPE:
+ Optional[Dict[str, str]]
+
+
+ DEFAULT:
+ dict()
+
+
Optional[str]
+ str
DEFAULT:
@@ -2533,7 +2533,7 @@ Apache StreamPipes meets Python! We are working highly motivated on a Python library to interact with StreamPipes. In this way, we would like to unite the power of StreamPipes to easily connect to and read from different data sources, especially in the IoT domain, and the amazing universe of data analytics libraries in Python.
"},{"location":"#quickstart","title":"\u26a1\ufe0f Quickstart","text":"As a quick example, we demonstrate how to set up and configure a StreamPipes client. In addition, we will get the available data lake measures out of StreamPipes.
from streampipes.client import StreamPipesClient\nfrom streampipes.client.config import StreamPipesClientConfig\nfrom streampipes.client.credential_provider import StreamPipesApiKeyCredentials\n\nconfig = StreamPipesClientConfig(\n credential_provider = StreamPipesApiKeyCredentials(\n username = \"test@streampipes.apache.org\",\n api_key = \"DEMO-KEY\",\n ),\n host_address = \"localhost\",\n https_disabled = True,\n port = 80\n)\n\nclient = StreamPipesClient(client_config=config)\n\n# get all available datat lake measures\nmeasures = client.dataLakeMeasureApi.all()\n\n# get amount of retrieved measures\nlen(measures)\n1\n# inspect the data lake measures as pandas dataframe\nmeasures.to_pandas()\nOutput:
measure_name timestamp_field ... pipeline_is_running num_event_properties\n0 test s0::timestamp ... False 2\n[1 rows x 6 columns]\nfrom streampipes.client.credential_provider import StreamPipesApiKeyCredentials\n\nStreamPipesApiKeyCredentials()\nSP_API_KEY and SP_USERNAME username is always the username that is used to log in into StreamPipes.
The api_key can be generated within the UI as demonstrated below:
This document describes how to easily set up your local dev environment to work on StreamPipes Python \ud83d\udc0d.
"},{"location":"getting-started/developing/#first-steps","title":"\ud83d\ude80 First Steps","text":"1) Set up your Python environment
Create a virtual Python environment using a tool of your choice. To manage dependencies, we use Poetry, so please install poetry in your local environment, e.g. via
pip install poetry\nOnce poetry is installed you can simply finalize your Python environment by running:
poetry install --with dev,stubs # install everything that is required for the development\npoetry install --with docs # install everything to work with the documentation\npoetry install --with dev,stubs,docs # install all optional dependencies related to development\n2) Install pre-commit hook
The pre-commit hook is run before every commit and takes care about code style, linting, type hints, import sorting, etc. It will stop your commit in case the changes do not apply the expected format. Always check to have the recent version of the pre-commit hook installed otherwise the CI build might fail. If you are interested, you can have a deeper look on the underlying library: pre-commit.
pre-commit install\nBelow we list some conventions that we have agreed on for creating StreamPipes Python. Please comply to them when you plan to contribute to this project. If you have any other suggestions or would like to discuss them, we would be happy to hear from you on our mailing list dev@streampipes.apache.org or in our discussions on GitHub.
1) Use numpy style for Python docstrings \ud83d\udcc4 Please stick to the numpy style when writing docstrings, as we require this for generating our documentation.
2) Provide tests \u2705 We are aiming for broad test coverage for the Python package and have therefore set a requirement of at least 90% unit test coverage. Therefore, please remember to write (unit) tests already during development. If you have problems with writing tests, don't hesitate to ask us for help directly in the PR or even before that via our mailing list (see above).
3) Build a similar API as the Java client provides \ud83d\udd04 Whenever possible, please try to develop the API of the Python library the same as the Java client or Java SDK. By doing so, we would like to provide a consistent developer experience and the basis for automated testing in the future.
"},{"location":"getting-started/developing/#dependency-management","title":"\ud83d\udce6 Dependency Management","text":"In case you want to add a new dependency to StreamPipes you can use the following command:
poetry add <dep-name>\nIf the dependency is only required for development purpose or the documentation, please stick to one the following:
poetry add <dep-name> --group dev\npoetry add <dep-name> --group stubs\npoetry add <dep-name> --group docs\nIn case you want to regenerate the poetry lock file, e.g., in case you manually updated the pyproject.toml, the following command should be used:
poetry lock --no-update\nAfter that, you should install the current version of the poetry lock file to keep your local environment consistent (see command above.)
"},{"location":"getting-started/developing/#documentation","title":"\ud83d\udcdaDocumentation","text":"To build our documentation, we use Materials for MkDocs. All files can be found within the docs directory. To pre-view your local version of the documentation, you can use the following command:
make livedoc\nBroadly speaking, we plan to expand or add new aspects/functionality to the library where we are focusing on the following:
In case you want to have a more detailed look on what we are currently planning, have a look at our open issues(more short-term driven).
Of course, contributions are always highly appreciated \ud83d\udd2e
Stay tuned!
"},{"location":"getting-started/developing/#contributing","title":"\ud83d\udc68\u200d\ud83d\udcbb Contributing","text":"Before opening a pull request, review the Get Involved page. It lists information that is required for contributing to StreamPipes.
When you contribute code, you affirm that the contribution is your original work and that you license the work to the project under the project's open source license. Whether or not you state this explicitly, by submitting any copyrighted material via pull request, email, or other means you agree to license the material under the project's open source license and warrant that you have the legal authority to do so.
"},{"location":"getting-started/first-steps/","title":"First Steps","text":""},{"location":"getting-started/first-steps/#installation","title":"\ud83d\udcda Installation","text":"The StreamPipes Python library is meant to work with Python 3.8 and above. Installation can be done via pip: You can install the latest development version from GitHub, as so:
pip install streampipes\n\n# if you want to have the current development state you can also execute\npip install git+https://github.com/apache/streampipes.git#subdirectory=streampipes-client-python\n# the corresponding documentation can be found here: https://streampipes.apache.org/docs/docs/python/dev/\nWhen working with the StreamPipes Python library it is inevitable to have a running StreamPipes instance to connect and interact with. In case you don't have a running instance at hand, you can easily set up one on your local machine. Hereby you need to consider that StreamPipes supports different message broker (e.g., Kafka, NATS). We will demonstrate below how you can easily set up StreamPipes for both supported message brokers.
"},{"location":"getting-started/first-steps/#start-streampipes-via-docker-compose","title":"\ud83d\udc33 Start StreamPipes via Docker Compose","text":"The easiest and therefore recommend way to get StreamPipes started is by using docker compose. Therefore, you need Docker running. You can check if Docker is ready on your machine by executing.
docker ps\nCONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES\n... ... ... ... ... ... ...\ndocker compose here."},{"location":"getting-started/first-steps/#setup-streampipes-with-nats-as-message-broker","title":"Setup StreamPipes with NATS as message broker","text":"The following shows how you can set up a StreamPipes instance that uses NATS as messaging layer. So in this scenario, we will go with docker-compose.nats.yml. Thereby, when running locally, we need to add the following port mapping entry to services.nats.ports:
- 4222:4222\nAfter this modification is applied, StreamPipes can simply be started with this command:
docker-compose -f docker-compose.nats.yml up -d\nOnce all services are started, you can access StreamPipes via http://localhost.
Alternatively, you can use docker-compose.yml to start StreamPipes with Kafka as messaging layer. When running locally we have to modify services.kafka.environment and add the ports to services.kafka.ports:
environment:\nKAFKA_LISTENER_SECURITY_PROTOCOL_MAP: PLAINTEXT:PLAINTEXT,OUTSIDE:PLAINTEXT\nKAFKA_ADVERTISED_LISTENERS: PLAINTEXT://:9092,OUTSIDE://localhost:9094\nKAFKA_LISTENERS: PLAINTEXT://:9092,OUTSIDE://:9094\n...\nports:\n- 9094:9094\ndocker-compose -f docker-compose.yml up -d\nOnce all services are started, you can access StreamPipes via http://localhost.
In case you want to have more control over your StreamPipes setup, you might take a look at our deployment CLI.
Have fun discovering StreamPipes and our Python library \ud83d\ude80
"},{"location":"getting-started/quickstart/","title":"Quickstart","text":""},{"location":"getting-started/quickstart/#quickstart","title":"\u26a1\ufe0f Quickstart","text":"As a quick example, we demonstrate how to set up and configure a StreamPipes client. In addition, we will get the available data lake measures out of StreamPipes.
from streampipes.client import StreamPipesClient\nfrom streampipes.client.config import StreamPipesClientConfig\nfrom streampipes.client.credential_provider import StreamPipesApiKeyCredentials\n\nconfig = StreamPipesClientConfig(\n credential_provider = StreamPipesApiKeyCredentials(\n username = \"test@streampipes.apache.org\",\n api_key = \"DEMO-KEY\",\n ),\n host_address = \"localhost\",\n https_disabled = True,\n port = 80\n)\n\nclient = StreamPipesClient(client_config=config)\n\n# get all available datat lake measures\nmeasures = client.dataLakeMeasureApi.all()\n\n# get amount of retrieved measures\nlen(measures)\n1\n# inspect the data lake measures as pandas dataframe\nmeasures.to_pandas()\nOutput:
measure_name timestamp_field ... pipeline_is_running num_event_properties\n0 test s0::timestamp ... False 2\n[1 rows x 6 columns]\nfrom streampipes.client.credential_provider import StreamPipesApiKeyCredentials\n\nStreamPipesApiKeyCredentials()\nSP_API_KEY and SP_USERNAME username is always the username that is used to log in into StreamPipes. The api_key can be generated within the UI as demonstrated below:
Implementation of the StreamPipes client. The client is designed as the central point of interaction with the StreamPipes API and provides all functionalities to communicate with the API.
"},{"location":"reference/client/client/#streampipes.client.client.StreamPipesClient","title":"StreamPipesClient(client_config, logging_level=logging.INFO)","text":"The client to connect to StreamPipes.
This is the central point of contact with StreamPipes and provides all the functionalities to interact with it.
The client provides so-called \"endpoints\" each of which refers to an endpoint of the StreamPipes API, e.g. .dataLakeMeasureApi. An endpoint provides the actual methods to interact with StreamPipes API.
client_config Configures the client to connect properly to the StreamPipes instance.
TYPE: StreamPipesClientConfig
logging_level Influences the log messages emitted by the StreamPipesClient
TYPE: Optional[int] DEFAULT: INFO
dataLakeMeasureApi Instance of the data lake measure endpoint
TYPE: DataLakeMeasureEndpoint
dataStreamApi Instance of the data stream endpoint
TYPE: DataStreamEndpoint
AttributeError: In case an invalid configuration of the StreamPipesClientConfig is passed
Examples:
from streampipes.client import StreamPipesClient\nfrom streampipes.client.config import StreamPipesClientConfig\nfrom streampipes.client.credential_provider import StreamPipesApiKeyCredentials\nclient_config = StreamPipesClientConfig(\n credential_provider=StreamPipesApiKeyCredentials(\n username=\"test-user\",\n api_key=\"api-key\"\n ),\n host_address=\"localhost\",\n https_disabled=True\n)\nThe following way of instantiating a client instance is intended to be consistent with the StreamPipes Java client.
client = StreamPipesClient.create(client_config=client_config)\nIf you prefer a more pythonic way, you can simply write:
client = StreamPipesClient(client_config=client_config)\nTo interact with an endpoint:
data_lake_measures = client.dataLakeMeasureApi.all()\nTo inspect returned data as a pandas dataframe:
data_lake_measures.to_pandas()\n#\n# measure_name timestamp_field ... pipeline_is_running num_event_properties\n# 0 test s0::timestamp ... False 2\n# [1 rows x 6 columns]\nbase_api_path: str property","text":"Constructs the basic API URL from the given client_config.
base_api_path basic API path of the connected StreamPipes instance
TYPE: str
http_headers: Dict[str, str] property","text":"Returns the HTTP headers used for all requests.
The HTTP headers are composed of the authentication headers supplied by the credential provider and additional required headers (currently this is only the application header).
RETURNS DESCRIPTIONhttp_headers Header information for HTTP requests as string key-value pairs.
TYPE: Dict[str, str]
create(client_config, logging_level=logging.INFO) classmethod","text":"Returns an instance of the StreamPipesPythonClient.
Provides consistency to the StreamPipes Java client.
PARAMETER DESCRIPTIONclient_config Configures the client to connect properly to the StreamPipes instance.
TYPE: StreamPipesClientConfig
logging_level Influences the log messages emitted by the StreamPipesClient.
TYPE: int DEFAULT: INFO
StreamPipesClient"},{"location":"reference/client/client/#streampipes.client.client.StreamPipesClient.describe","title":"describe()","text":"Prints a short description of the connected StreamPipes instance and the available resources to the console.
RETURNS DESCRIPTION None Examples:
client.describe()\nHi there!\nYou are connected to a StreamPipes instance running at http://localhost:80.\nThe following StreamPipes resources are available with this client:\n6x DataStreams\n1x DataLakeMeasures\nConfiguration class for the StreamPipes client.
"},{"location":"reference/client/config/#streampipes.client.config.StreamPipesClientConfig","title":"StreamPipesClientConfig(credential_provider, host_address, https_disabled=False, port=80) dataclass","text":"Configure the StreamPipes client in accordance to the actual StreamPipes instance to connect to.
An instance is provided to the StreamPipesClient to configure it properly.
credential_provider Provides the credentials to authenticate with the StreamPipes API.
TYPE: CredentialProvider
host_address Host address of the StreamPipes instance to connect to. Should be provided without the protocol/scheme, e.g. as localhost or streampipes.xyz.
TYPE: str
https_disabled Determines whether https is used to connect to StreamPipes.
TYPE: Optional[bool] DEFAULT: False
port Specifies the port under which the StreamPipes API is available, e.g., 80 (with http) or 443 (with https)
TYPE: Optional[int] DEFAULT: 80
Examples:
see StreamPipesClient
"},{"location":"reference/client/credential_provider/","title":"Credential provider","text":"Implementation of credential providers. A credential provider supplies the specified sort of credentials in the appropriate HTTP header format. The headers are then used by the client to connect to StreamPipes.
"},{"location":"reference/client/credential_provider/#streampipes.client.credential_provider.CredentialProvider","title":"CredentialProvider","text":" Bases: ABC
Abstract implementation of a credential provider. Must be inherited by all credential providers.
"},{"location":"reference/client/credential_provider/#streampipes.client.credential_provider.CredentialProvider.make_headers","title":"make_headers(http_headers=None)","text":"Creates the HTTP headers for the specific credential provider.
Concrete authentication headers must be defined in the implementation of a credential provider.
PARAMETER DESCRIPTIONhttp_headers Additional HTTP headers the generated headers are extended by.
TYPE: Optional[Dict[str, str]] DEFAULT: None
https_headers Dictionary with header information as string key-value pairs. Contains all pairs given as parameter plus the header pairs for authentication determined by the credential provider.
TYPE: Dict[str, str]
StreamPipesApiKeyCredentials(username=None, api_key=None)","text":" Bases: CredentialProvider
A credential provider that allows authentication via a StreamPipes API Token.
The required token can be generated via the StreamPipes UI (see the description on our start-page.
Both parameters can either be passed as arguments or remain unset. If they are not passed, they are retrieved from environment variables:
SP_USERNAME is expected to contain the usernameSP_API_KEY is expected to contain the API keyusername The username to which the API token is granted, e.g., demo-user@streampipes.apche.org. If not passed, the username is retrieved from environment variable SP_USERNAME.
TYPE: Optional[str] DEFAULT: None
api_key The StreamPipes API key as it is displayed in the UI. If not passed, the api key is retrieved from environment variable SP_API_KEY
TYPE: Optional[str] DEFAULT: None
Examples:
see StreamPipesClient
"},{"location":"reference/client/credential_provider/#streampipes.client.credential_provider.StreamPipesApiKeyCredentials.from_env","title":"from_env(username_env, api_key_env) classmethod","text":"DEPRECATED - use the class constructor instead
Returns an API key provider parameterized via environment variables.
PARAMETER DESCRIPTIONusername_env Name of the environment variable that contains the username
TYPE: str
api_key_env Name of the environment variable that contains the API key
TYPE: str
StreamPipesApiKeyCredentials RAISES DESCRIPTION KeyError If one of the environment variables is not defined
"},{"location":"reference/client/credential_provider/#streampipes.client.credential_provider.StreamPipesApiKeyCredentials.make_headers","title":"make_headers(http_headers=None)","text":"Creates the HTTP headers for the specific credential provider.
Concrete authentication headers must be defined in the implementation of a credential provider.
PARAMETER DESCRIPTIONhttp_headers Additional HTTP headers the generated headers are extended by.
TYPE: Optional[Dict[str, str]] DEFAULT: None
https_headers Dictionary with header information as string key-value pairs. Contains all pairs given as parameter plus the header pairs for authentication determined by the credential provider.
TYPE: Dict[str, str]
General implementation for an endpoint. Provided classes and assets are aimed to be used for developing endpoints. An endpoint provides all options to communicate with a dedicated part of StreamPipes in a handy way.
"},{"location":"reference/endpoint/endpoint/#streampipes.endpoint.endpoint.APIEndpoint","title":"APIEndpoint(parent_client)","text":" Bases: Endpoint
Abstract implementation of an API endpoint.
Serves as template for all endpoints of the StreamPipes API. By design, endpoints are only instantiated within the __init__ method of the StreamPipesClient.
all()","text":"Get all resources of this endpoint provided by the StreamPipes API.
Results are provided as an instance of a ResourceContainer that allows to handle the returned resources in a comfortable and pythonic way.
container Container element that bundles the returned resources
TYPE: ResourceContainer
build_url()","text":"Builds the endpoint's URL of the API path.
RETURNS DESCRIPTIONurl The URL of the endpoint
TYPE: str
get(identifier, **kwargs)","text":"Queries the specified resource from the API endpoint.
PARAMETER DESCRIPTIONidentifier The identifier of the resource to be queried.
TYPE: str
resource The specified resource as an instance of the corresponding model class.
TYPE: Resource
post(resource)","text":"Allows to post a resource to the StreamPipes API.
PARAMETER DESCRIPTIONresource The resource to be posted.
TYPE: Resource
None"},{"location":"reference/endpoint/endpoint/#streampipes.endpoint.endpoint.Endpoint","title":"Endpoint(parent_client)","text":" Bases: ABC
Abstract implementation of a StreamPipes endpoint.
Serves as template for all endpoints used for interaction with a StreamPipes instance. By design, endpoints are only instantiated within the __init__ method of the StreamPipesClient.
parent_client This parameter expects the instance of StreamPipesClient the endpoint is attached to.
TYPE: StreamPipesClient
MessagingEndpoint(parent_client)","text":" Bases: Endpoint
Abstract implementation of a StreamPipes messaging endpoint.
Serves as template for all endpoints used for interacting with the StreamPipes messaging layer directly. Therefore, they need to provide the functionality to talk with the broker system running in StreamPipes. By design, endpoints are only instantiated within the __init__ method of the StreamPipesClient.
broker: Broker property writable","text":"Defines the broker instance that is used to connect to StreamPipes' messaging layer.
This instance enables the client to authenticate to the broker used in the target StreamPipes instance, to consume messages from and to write messages to the broker.
RAISES DESCRIPTIONMessagingEndpointNotConfiguredError If the endpoint is used before the broker instance is set via configure()
broker The broker instance to be used to communicate with StreamPipes' messaging layer.
TYPE: Broker
configure(broker)","text":"Configures the message endpoint by setting the broker instance to be used.
This configuration step is required before the endpoint can be actually used. The based broker instance is passed to an internal property.
broker Broker instance that should be used for this endpoint
TYPE: Broker
None"},{"location":"reference/endpoint/exceptions/","title":"Exceptions","text":"Custom exceptions dedicated to the endpoints module.
"},{"location":"reference/endpoint/exceptions/#streampipes.endpoint.exceptions.MessagingEndpointNotConfiguredError","title":"MessagingEndpointNotConfiguredError(endpoint_name)","text":" Bases: Exception
Exception that indicates that an instance of a messaging endpoint has not been configured.
This error occurs when an instance of a messaging endpoint is used before the broker instance to be used is configured by passing it to the configure() method.
endpoint_name The name of the endpoint that caused the error
TYPE: str
Specific implementation of the StreamPipes API's data lake measure endpoints. This endpoint allows to consume data stored in StreamPipes' data lake.
"},{"location":"reference/endpoint/api/data_lake_measure/#streampipes.endpoint.api.data_lake_measure.DataLakeMeasureEndpoint","title":"DataLakeMeasureEndpoint(parent_client)","text":" Bases: APIEndpoint
Implementation of the DataLakeMeasure endpoint.
This endpoint provides an interface to all data stored in the StreamPipes data lake.
Consequently, it allows querying metadata about available data sets (see all() method). The metadata is returned as an instance of DataLakeMeasures.
In addition, the endpoint provides direct access to the data stored in the data laka by querying a specific data lake measure using the get() method.
Examples:
from streampipes.client import StreamPipesClient\nfrom streampipes.client.config import StreamPipesClientConfig\nfrom streampipes.client.credential_provider import StreamPipesApiKeyCredentials\nclient_config = StreamPipesClientConfig(\n credential_provider=StreamPipesApiKeyCredentials(username=\"test-user\", api_key=\"api-key\"),\n host_address=\"localhost\",\n port=8082,\n https_disabled=True\n)\nclient = StreamPipesClient.create(client_config=client_config)\n# get all existing data lake measures from StreamPipes\ndata_lake_measures = client.dataLakeMeasureApi.all()\n\n# let's take a look how many we got\nlen(data_lake_measures)\n5\n# Retrieve a specific data lake measure as a pandas DataFrame\nflow_rate_pd = client.dataLakeMeasureApi.get(identifier=\"flow-rate\").to_pandas()\nflow_rate_pd\n time density mass_flow sensorId sensor_fault_flags temperature volume_flow\n0 2023-02-24T16:19:41.472Z 50.872730 3.309556 flowrate02 False 44.448483 5.793138\n1 2023-02-24T16:19:41.482Z 47.186588 5.608580 flowrate02 False 40.322033 0.058015\n2 2023-02-24T16:19:41.493Z 46.735321 7.692881 flowrate02 False 49.239639 10.283526\n3 2023-02-24T16:19:41.503Z 40.169796 3.632898 flowrate02 False 49.933754 6.893441\n4 2023-02-24T16:19:41.513Z 49.635124 0.711260 flowrate02 False 50.106617 2.999871\n.. ... ... ... ... ... ... ...\n995 2023-02-24T16:19:52.927Z 50.057495 1.740114 flowrate02 False 46.558231 1.818237\n996 2023-02-24T16:19:52.94Z 41.038895 7.211723 flowrate02 False 48.048622 2.127493\n997 2023-02-24T16:19:52.952Z 45.837013 7.770180 flowrate02 False 48.188026 7.892062\n998 2023-02-24T16:19:52.965Z 43.389065 4.458602 flowrate02 False 48.280899 5.733892\n999 2023-02-24T16:19:52.977Z 44.056030 2.592060 flowrate02 False 47.505951 4.260697\nAs you can see, the returned amount of rows per default is 1000. We can modify this behavior by passing the limit paramter.
flow_rate_pd = client.dataLakeMeasureApi.get(identifier=\"flow-rate\", limit=10).to_pandas()\nlen(flow_rate_pd)\n10\nIf we are only interested in the values for density, columns allows us to select the columns to be returned:
flow_rate_pd = client.dataLakeMeasureApi.get(identifier=\"flow-rate\", columns='density', limit=3).to_pandas()\nflow_rate_pd\n time density\n0 2023-02-24T16:19:41.472Z 50.872730\n1 2023-02-24T16:19:41.482Z 47.186588\n2 2023-02-24T16:19:41.493Z 46.735321\nThis is only a subset of the available query parameters, find them at MeasurementGetQueryConfig.
"},{"location":"reference/endpoint/api/data_lake_measure/#streampipes.endpoint.api.data_lake_measure.DataLakeMeasureEndpoint.all","title":"all()","text":"Get all resources of this endpoint provided by the StreamPipes API.
Results are provided as an instance of a ResourceContainer that allows to handle the returned resources in a comfortable and pythonic way.
container Container element that bundles the returned resources
TYPE: ResourceContainer
build_url()","text":"Builds the endpoint's URL of the API path.
RETURNS DESCRIPTIONurl The URL of the endpoint
TYPE: str
get(identifier, **kwargs)","text":"Queries the specified data lake measure from the API.
By default, the maximum number of returned records is 1000. This behaviour can be influenced by passing the parameter limit with a different value (see MeasurementGetQueryConfig).
identifier The identifier of the data lake measure to be queried.
TYPE: str
**kwargs Keyword arguments can be used to provide additional query parameters. The available query parameters are defined by the MeasurementGetQueryConfig.
TYPE: Optional[Dict[str, Any]] DEFAULT: {}
measurement The specified data lake measure
TYPE: QueryResult
Examples:
see directly at DataLakeMeasureEndpoint.
"},{"location":"reference/endpoint/api/data_lake_measure/#streampipes.endpoint.api.data_lake_measure.DataLakeMeasureEndpoint.post","title":"post(resource)","text":"Allows to post a resource to the StreamPipes API.
PARAMETER DESCRIPTIONresource The resource to be posted.
TYPE: Resource
None"},{"location":"reference/endpoint/api/data_lake_measure/#streampipes.endpoint.api.data_lake_measure.MeasurementGetQueryConfig","title":"MeasurementGetQueryConfig","text":" Bases: BaseModel
Config class describing the parameters of the get() method for measurements.
This config class is used to validate the provided query parameters for the GET endpoint of measurements. Additionally, it takes care of the conversion to a proper HTTP query string. Thereby, parameter names are adapted to the naming of the StreamPipes API, for which Pydantic aliases are used.
ATTRIBUTE DESCRIPTIONcolumns A comma separated list of column names (e.g., time,value) If provided, the returned data only consists of the given columns. Please be aware that the column time as an index is always included.
TYPE: Optional[List[str]]
end_date Limits the queried data to only include data that is older than the specified time. In other words, any data that occurred after the end_date will not be included in the query results.
TYPE: Optional[datetime]
limit Amount of records returned at maximum (default: 1000) This needs to be at least 1
TYPE: Optional[int]
offset Offset to be applied to returned data This needs to be at least 0
TYPE: Optional[int]
order Ordering of query results Allowed values: ASC and DESC (default: ASC)
TYPE: Optional[str]
page_no Page number used for paging operation This needs to be at least 1
TYPE: Optional[int]
start_date Limits the queried data to only include data that is newer than the specified time. In other words, any data that occurred before the start_date will not be included in the query results.
TYPE: Optional[datetime]
Config","text":"Pydantic Config class
"},{"location":"reference/endpoint/api/data_lake_measure/#streampipes.endpoint.api.data_lake_measure.MeasurementGetQueryConfig.build_query_string","title":"build_query_string()","text":"Builds a HTTP query string for the config.
This method returns an HTTP query string for the invoking config. It follows the following structure ?param1=value1¶m2=value2.... This query string is not an entire URL, instead it needs to appended to an API path.
query_param_string HTTP query params string (?param1=value1¶m2=value2...)
TYPE: str
StreamPipesQueryValidationError","text":" Bases: Exception
A custom exception to be raised when the validation of query parameter causes an error.
"},{"location":"reference/endpoint/api/data_stream/","title":"Data stream","text":"Specific implementation of the StreamPipes API's data stream endpoints.
"},{"location":"reference/endpoint/api/data_stream/#streampipes.endpoint.api.data_stream.DataStreamEndpoint","title":"DataStreamEndpoint(parent_client)","text":" Bases: APIEndpoint
Implementation of the DataStream endpoint.
Consequently, it allows querying metadata about available data streams (see all() method). The metadata is returned as an instance of DataStreams.
Examples:
from streampipes.client import StreamPipesClient\nfrom streampipes.client.config import StreamPipesClientConfig\nfrom streampipes.client.credential_provider import StreamPipesApiKeyCredentials\n\nclient_config = StreamPipesClientConfig(\n credential_provider=StreamPipesApiKeyCredentials(username=\"test-user\", api_key=\"api-key\"),\n host_address=\"localhost\",\n port=8082,\n https_disabled=True\n)\nclient = StreamPipesClient.create(client_config=client_config)\n# let's get all existing data streams in StreamPipes\ndata_streams = client.dataStreamApi.all()\nlen(data_streams)\n2\nall()","text":"Get all resources of this endpoint provided by the StreamPipes API.
Results are provided as an instance of a ResourceContainer that allows to handle the returned resources in a comfortable and pythonic way.
container Container element that bundles the returned resources
TYPE: ResourceContainer
build_url()","text":"Builds the endpoint's URL of the API path.
RETURNS DESCRIPTIONurl The URL of the endpoint
TYPE: str
get(identifier, **kwargs)","text":"Queries the specified resource from the API endpoint.
PARAMETER DESCRIPTIONidentifier The identifier of the resource to be queried.
TYPE: str
resource The specified resource as an instance of the corresponding model class.
TYPE: Resource
post(resource)","text":"Allows to post a resource to the StreamPipes API.
PARAMETER DESCRIPTIONresource The resource to be posted.
TYPE: Resource
None"},{"location":"reference/endpoint/api/version/","title":"Version","text":"Specific implementation of the StreamPipes API's version endpoint.
"},{"location":"reference/endpoint/api/version/#streampipes.endpoint.api.version.VersionEndpoint","title":"VersionEndpoint(parent_client)","text":" Bases: APIEndpoint
Implementation of the Versions endpoint.
This endpoint provides metadata about the StreamPipes version of the connected instance. It only allows to apply the get() method with an empty string as identifier.
parent_client The instance of StreamPipesClient the endpoint is attached to.
TYPE: StreamPipesClient
Examples:
>>> from streampipes.client import StreamPipesClient\n>>> from streampipes.client.config import StreamPipesClientConfig\n>>> from streampipes.client.credential_provider import StreamPipesApiKeyCredentials\n>>> client_config = StreamPipesClientConfig(\n... credential_provider=StreamPipesApiKeyCredentials(username=\"test-user\", api_key=\"api-key\"),\n... host_address=\"localhost\",\n... port=8082,\n... https_disabled=True\n... )\n>>> client = StreamPipesClient.create(client_config=client_config)\n>>> client.versionApi.get(identifier=\"\").to_dict(use_source_names=False)\n{'backend_version': '0.92.0-SNAPSHOT'}\nall()","text":"Usually, this method returns information about all resources provided by this endpoint. However, this endpoint does not support this kind of operation.
RAISES DESCRIPTIONNotImplementedError This endpoint does not return multiple entries, therefore this method is not available.
"},{"location":"reference/endpoint/api/version/#streampipes.endpoint.api.version.VersionEndpoint.build_url","title":"build_url()","text":"Builds the endpoint's URL of the API path.
RETURNS DESCRIPTIONurl The URL of the endpoint
TYPE: str
get(identifier, **kwargs)","text":"Queries the resource from the API endpoint.
For this endpoint only one resource is available.
PARAMETER DESCRIPTIONidentifier Not supported by this endpoint, is set to an empty string.
TYPE: str
ValueError Non-empty identifier is not supported by this endpoint. Please set identifier to an empty string or None.
versions The specified resource as an instance of the corresponding model class(Version). # noqa: 501
TYPE: Version
post(resource)","text":"Usually, this method allows to create via this endpoint. Since the data represented by this endpoint is immutable, it does not support this kind of operation.
RAISES DESCRIPTIONNotImplementedError This endpoint does not allow for POST requests, therefore this method is not available.
"},{"location":"reference/function_zoo/river_function/","title":"River function","text":""},{"location":"reference/function_zoo/river_function/#streampipes.function_zoo.river_function.OnlineML","title":"OnlineML(client, stream_ids, model, prediction_type=RuntimeType.STRING.value, supervised=False, target_label=None, on_start=lambda self, context: None, on_event=lambda self, event, streamId: None, on_stop=lambda self: None)","text":"Wrapper class to enable an easy usage for Online Machine Learning models of the River library.
It creates a StreamPipesFunction to train a model with the incoming events of a data stream and creates an output data stream that publishes the prediction to StreamPipes.
PARAMETER DESCRIPTIONclient The client for the StreamPipes API.
TYPE: StreamPipesClient
stream_ids The ids of the data stream to train the model.
TYPE: List[str]
model The model to train. It meant to be a River model/pipeline, but can be every model with a 'learn_one' and 'predict_one' methode.
TYPE: Any
prediction_type The data type of the prediction. Is only needed when you continue to work with the prediction in StreamPipes.
TYPE: str DEFAULT: value
supervised Define if the model is supervised or unsupervised.
TYPE: bool DEFAULT: False
target_label Define the name of the target attribute if the model is supervised.
TYPE: Optional[str] DEFAULT: None
on_start A function to be called when this StreamPipesFunction gets started.
TYPE: Callable[[Any, FunctionContext], None] DEFAULT: lambda self, context: None
on_event A function to be called when this StreamPipesFunction receives an event.
TYPE: Callable[[Any, Dict[str, Any], str], None] DEFAULT: lambda self, event, streamId: None
on_stop A function to be called when this StreamPipesFunction gets stopped.
TYPE: Callable[[Any], None] DEFAULT: lambda self: None
set_learning(learning)","text":"Start or stop the training of the model.
PARAMETER DESCRIPTIONlearning Defines if the training should be continued
TYPE: bool
start()","text":"Registers the function and starts the training.
"},{"location":"reference/function_zoo/river_function/#streampipes.function_zoo.river_function.OnlineML.stop","title":"stop()","text":"Stops the function and ends the training forever.
"},{"location":"reference/function_zoo/river_function/#streampipes.function_zoo.river_function.RiverFunction","title":"RiverFunction(function_definition, model, supervised, target_label, on_start, on_event, on_stop)","text":" Bases: StreamPipesFunction
Implementation of a StreamPipesFunction to enable an easy usage for Online Machine Learning models of the River library.
The function trains the model with the incoming events and publishes the prediction to an output data stream.
PARAMETER DESCRIPTIONfunction_definition The function definition which contains the output stream.
TYPE: FunctionDefinition
model The model to train. It meant to be a River model/pipeline, but can be every model with a 'learn_one' and 'predict_one' method.
TYPE: Any
supervised Define if the model is supervised or unsupervised.
TYPE: bool
target_label Define the name of the target attribute if the model is supervised.
TYPE: Optional[str]
on_start A function to be called when this StreamPipesFunction gets started.
TYPE: Callable[[Any, FunctionContext], None]
on_event A function to be called when this StreamPipesFunction receives an event.
TYPE: Callable[[Any, Dict[str, Any], str], None]
on_stop A function to be called when this StreamPipesFunction gets stopped.
TYPE: Callable[[Any], None]
add_output(stream_id, event)","text":"Send an event via an output data stream to StreamPipes.
PARAMETER DESCRIPTIONstream_id The id of the output data stream
TYPE: str
event The event which should be sent
TYPE: Dict[str, Any]
None"},{"location":"reference/function_zoo/river_function/#streampipes.function_zoo.river_function.RiverFunction.getFunctionId","title":"getFunctionId()","text":"Returns the id of the function.
RETURNS DESCRIPTIONfunction_id Identification object of the StreamPipes function
TYPE: FunctionId
onEvent(event, streamId)","text":"Trains the model with the incoming events and sends the prediction back to StreamPipes.
PARAMETER DESCRIPTIONevent The incoming event that serves as input for the function
TYPE: Dict[str, Any]
streamId Identifier of the corresponding data stream
TYPE: str
None"},{"location":"reference/function_zoo/river_function/#streampipes.function_zoo.river_function.RiverFunction.onServiceStarted","title":"onServiceStarted(context)","text":"Executes the on_start method of the function.
context The functions' context
TYPE: FunctionContext
None"},{"location":"reference/function_zoo/river_function/#streampipes.function_zoo.river_function.RiverFunction.onServiceStopped","title":"onServiceStopped()","text":"Executes the on_stop function.
requiredStreamIds()","text":"Get the ids of the streams needed by the function.
RETURNS DESCRIPTIONstream_ids List of the stream ids
TYPE: List[str]
stop()","text":"Stops the function and disconnects from the output streams.
"},{"location":"reference/functions/function_handler/","title":"Function handler","text":""},{"location":"reference/functions/function_handler/#streampipes.functions.function_handler.FunctionHandler","title":"FunctionHandler(registration, client)","text":"The function handler manages the StreamPipes Functions.
It controls the connection to the brokers, starts the functions, manages the broadcast of the live data and is able to stop the connection to the brokers and functions.
PARAMETER DESCRIPTIONregistration The registration, that contains the StreamPipesFunctions.
TYPE: Registration
client The client to interact with the API.
TYPE: StreamPipesClient
stream_contexts Map of all data stream contexts
TYPE: Dict[str, DataStreamContext]
brokers List of all registered brokers
TYPE: List[Broker]
disconnect()","text":"Disconnects from the brokers and stops all functions.
RETURNS DESCRIPTIONNone"},{"location":"reference/functions/function_handler/#streampipes.functions.function_handler.FunctionHandler.force_stop_functions","title":"force_stop_functions()","text":"Stops the StreamPipesFunctions when the event loop was stopped without stopping the functions.
RETURNS DESCRIPTIONNone WARNS DESCRIPTION UserWarning If there is a running event loop and the functions should be stopped by disconnecting from the broker.
"},{"location":"reference/functions/function_handler/#streampipes.functions.function_handler.FunctionHandler.initializeFunctions","title":"initializeFunctions()","text":"Creates the context for every data stream and starts the event loop to manage the StreamPipes Functions.
RETURNS DESCRIPTIONNone"},{"location":"reference/functions/registration/","title":"Registration","text":""},{"location":"reference/functions/registration/#streampipes.functions.registration.Registration","title":"Registration()","text":"Manages the existing StreamPipesFunctions and registers them.
ATTRIBUTE DESCRIPTIONfunctions List of all registered StreamPipesFunction
TYPE: List[StreamPipesFunction]
getFunctions()","text":"Get all registered functions.
This method exists to be consistent with the Java client.
RETURNS DESCRIPTIONfunctions List of all registered functions.
TYPE: List[StreamPipesFunction]
register(streampipes_function)","text":"Registers a new function.
PARAMETER DESCRIPTIONstreampipes_function The function to register.
TYPE: StreamPipesFunction
self The updated Registration instance
TYPE: Registration
StreamPipesFunction(function_definition=None)","text":" Bases: ABC
Abstract implementation of a StreamPipesFunction.
A StreamPipesFunction allows users to get the data of a StreamPipes data streams easily. It makes it possible to work with the live data in python and enables to use the powerful data analytics libraries there.
PARAMETER DESCRIPTIONfunction_definition The definition of the function that contains metadata about the connected function
TYPE: Optional[FunctionDefinition] DEFAULT: None
output_collectors List of all output collectors which are created based on the provided function definitions.
TYPE: Dict[str, OutputCollector]
add_output(stream_id, event)","text":"Send an event via an output data stream to StreamPipes.
PARAMETER DESCRIPTIONstream_id The id of the output data stream
TYPE: str
event The event which should be sent
TYPE: Dict[str, Any]
None"},{"location":"reference/functions/streampipes_function/#streampipes.functions.streampipes_function.StreamPipesFunction.getFunctionId","title":"getFunctionId()","text":"Returns the id of the function.
RETURNS DESCRIPTIONfunction_id Identification object of the StreamPipes function
TYPE: FunctionId
onEvent(event, streamId) abstractmethod","text":"Is called for every event of a data stream.
PARAMETER DESCRIPTIONevent The received event from the data stream.
TYPE: Dict[str, Any]
streamId The id of the data stream which the event belongs to.
TYPE: str
None"},{"location":"reference/functions/streampipes_function/#streampipes.functions.streampipes_function.StreamPipesFunction.onServiceStarted","title":"onServiceStarted(context) abstractmethod","text":"Is called when the function is started.
PARAMETER DESCRIPTIONcontext The context in which the function is started.
TYPE: FunctionContext
None"},{"location":"reference/functions/streampipes_function/#streampipes.functions.streampipes_function.StreamPipesFunction.onServiceStopped","title":"onServiceStopped() abstractmethod","text":"Is called when the function is stopped.
RETURNS DESCRIPTIONNone"},{"location":"reference/functions/streampipes_function/#streampipes.functions.streampipes_function.StreamPipesFunction.requiredStreamIds","title":"requiredStreamIds()","text":"Get the ids of the streams needed by the function.
RETURNS DESCRIPTIONstream_ids List of the stream ids
TYPE: List[str]
stop()","text":"Stops the function and disconnects from the output streams.
"},{"location":"reference/functions/broker/broker/","title":"Broker","text":""},{"location":"reference/functions/broker/broker/#streampipes.functions.broker.broker.Broker","title":"Broker","text":" Bases: ABC
Abstract implementation of a broker for consumer and publisher.
It contains the basic logic to connect to a data stream.
"},{"location":"reference/functions/broker/broker/#streampipes.functions.broker.broker.Broker.connect","title":"connect(data_stream) async","text":"Connects to the broker running in StreamPipes.
PARAMETER DESCRIPTIONdata_stream Contains the meta information (resources) for a data stream.
TYPE: DataStream
None"},{"location":"reference/functions/broker/broker/#streampipes.functions.broker.broker.Broker.disconnect","title":"disconnect() abstractmethod async","text":"Closes the connection to the server.
RETURNS DESCRIPTIONNone"},{"location":"reference/functions/broker/broker_handler/","title":"Broker handler","text":""},{"location":"reference/functions/broker/broker_handler/#streampipes.functions.broker.broker_handler.SupportedBroker","title":"SupportedBroker","text":" Bases: Enum
Enum for the supported brokers.
"},{"location":"reference/functions/broker/broker_handler/#streampipes.functions.broker.broker_handler.UnsupportedBrokerError","title":"UnsupportedBrokerError(broker_name)","text":" Bases: Exception
Exception if a broker isn't implemented yet.
"},{"location":"reference/functions/broker/broker_handler/#streampipes.functions.broker.broker_handler.get_broker","title":"get_broker(data_stream, is_publisher=False)","text":"Derive the broker for the given data stream.
PARAMETER DESCRIPTIONdata_stream Data stream instance from which the broker is inferred
TYPE: DataStream
broker The corresponding broker instance derived from data stream.
TYPE: Broker
UnsupportedBrokerError Is raised when the given data stream belongs to a broker that is currently not supported by StreamPipes Python.
"},{"location":"reference/functions/broker/broker_handler/#streampipes.functions.broker.broker_handler.get_broker_description","title":"get_broker_description(data_stream)","text":"Derive the decription of the broker for the given data stream.
PARAMETER DESCRIPTIONdata_stream Data stream instance from which the broker is inferred
TYPE: DataStream
broker The corresponding broker description derived from data stream.
TYPE: SupportedBroker
UnsupportedBrokerError Is raised when the given data stream belongs to a broker that is currently not supported by StreamPipes Python.
"},{"location":"reference/functions/broker/consumer/","title":"Consumer","text":""},{"location":"reference/functions/broker/consumer/#streampipes.functions.broker.consumer.Consumer","title":"Consumer","text":" Bases: Broker
Abstract implementation a consumer for a broker.
A consumer allows to subscribe to a data stream.
"},{"location":"reference/functions/broker/consumer/#streampipes.functions.broker.consumer.Consumer.connect","title":"connect(data_stream) async","text":"Connects to the broker running in StreamPipes and creates a subscription.
PARAMETER DESCRIPTIONdata_stream Contains the meta information (resources) for a data stream.
TYPE: DataStream
None"},{"location":"reference/functions/broker/consumer/#streampipes.functions.broker.consumer.Consumer.disconnect","title":"disconnect() abstractmethod async","text":"Closes the connection to the server.
RETURNS DESCRIPTIONNone"},{"location":"reference/functions/broker/consumer/#streampipes.functions.broker.consumer.Consumer.get_message","title":"get_message() abstractmethod","text":"Get the published messages of the subscription.
RETURNS DESCRIPTIONiterator An async iterator for the messages.
TYPE: AsyncIterator
OutputCollector(data_stream)","text":"Collector for output events. The events are published to an output data stream. Therefore, the output collector establishes a connection to the broker.
PARAMETER DESCRIPTIONdata_stream The output data stream that will receive the events.
TYPE: DataStream
publisher The publisher instance that sends the data to StreamPipes
TYPE: Publisher
collect(event)","text":"Publishes an event to the output stream.
PARAMETER DESCRIPTIONevent The event to be published.
TYPE: Dict[str, Any]
None"},{"location":"reference/functions/broker/output_collector/#streampipes.functions.broker.output_collector.OutputCollector.disconnect","title":"disconnect()","text":"Disconnects the broker of the output collector.
RETURNS DESCRIPTIONNone"},{"location":"reference/functions/broker/publisher/","title":"Publisher","text":""},{"location":"reference/functions/broker/publisher/#streampipes.functions.broker.publisher.Publisher","title":"Publisher","text":" Bases: Broker
Abstract implementation of a publisher for a broker.
A publisher allows to publish events to a data stream.
"},{"location":"reference/functions/broker/publisher/#streampipes.functions.broker.publisher.Publisher.connect","title":"connect(data_stream) async","text":"Connects to the broker running in StreamPipes.
PARAMETER DESCRIPTIONdata_stream Contains the meta information (resources) for a data stream.
TYPE: DataStream
None"},{"location":"reference/functions/broker/publisher/#streampipes.functions.broker.publisher.Publisher.disconnect","title":"disconnect() abstractmethod async","text":"Closes the connection to the server.
RETURNS DESCRIPTIONNone"},{"location":"reference/functions/broker/publisher/#streampipes.functions.broker.publisher.Publisher.publish_event","title":"publish_event(event) abstractmethod async","text":"Publish an event to a connected data stream.
PARAMETER DESCRIPTIONevent The event to be published.
TYPE: Dict[str, Any]
None"},{"location":"reference/functions/broker/kafka/kafka_consumer/","title":"Kafka consumer","text":""},{"location":"reference/functions/broker/kafka/kafka_consumer/#streampipes.functions.broker.kafka.kafka_consumer.KafkaConsumer","title":"KafkaConsumer","text":" Bases: Consumer
Implementation of a consumer for Kafka
"},{"location":"reference/functions/broker/kafka/kafka_consumer/#streampipes.functions.broker.kafka.kafka_consumer.KafkaConsumer.connect","title":"connect(data_stream) async","text":"Connects to the broker running in StreamPipes and creates a subscription.
PARAMETER DESCRIPTIONdata_stream Contains the meta information (resources) for a data stream.
TYPE: DataStream
None"},{"location":"reference/functions/broker/kafka/kafka_consumer/#streampipes.functions.broker.kafka.kafka_consumer.KafkaConsumer.disconnect","title":"disconnect() async","text":"Closes the connection to the server.
RETURNS DESCRIPTIONNone"},{"location":"reference/functions/broker/kafka/kafka_consumer/#streampipes.functions.broker.kafka.kafka_consumer.KafkaConsumer.get_message","title":"get_message()","text":"Get the published messages of the subscription.
RETURNS DESCRIPTIONiterator An async iterator for the messages.
TYPE: AsyncIterator
KafkaMessage(data)","text":"An internal representation of a Kafka message
PARAMETER DESCRIPTIONdata The received Kafka message as byte array
"},{"location":"reference/functions/broker/kafka/kafka_message_fetcher/#streampipes.functions.broker.kafka.kafka_message_fetcher.KafkaMessageFetcher","title":"
KafkaMessageFetcher(consumer)","text":"Fetches the next message from Kafka
PARAMETER DESCRIPTIONconsumer The Kafka consumer
TYPE: Consumer
KafkaPublisher","text":" Bases: Publisher
Implementation of a publisher for Kafka
"},{"location":"reference/functions/broker/kafka/kafka_publisher/#streampipes.functions.broker.kafka.kafka_publisher.KafkaPublisher.connect","title":"connect(data_stream) async","text":"Connects to the broker running in StreamPipes.
PARAMETER DESCRIPTIONdata_stream Contains the meta information (resources) for a data stream.
TYPE: DataStream
None"},{"location":"reference/functions/broker/kafka/kafka_publisher/#streampipes.functions.broker.kafka.kafka_publisher.KafkaPublisher.disconnect","title":"disconnect() async","text":"Closes the connection to the server.
RETURNS DESCRIPTIONNone"},{"location":"reference/functions/broker/kafka/kafka_publisher/#streampipes.functions.broker.kafka.kafka_publisher.KafkaPublisher.publish_event","title":"publish_event(event) async","text":"Publish an event to a connected data stream.
PARAMETER DESCRIPTIONevent The event to be published.
TYPE: Dict[str, Any]
None"},{"location":"reference/functions/broker/nats/nats_consumer/","title":"Nats consumer","text":""},{"location":"reference/functions/broker/nats/nats_consumer/#streampipes.functions.broker.nats.nats_consumer.NatsConsumer","title":"NatsConsumer","text":" Bases: Consumer
Implementation of a consumer for NATS
"},{"location":"reference/functions/broker/nats/nats_consumer/#streampipes.functions.broker.nats.nats_consumer.NatsConsumer.connect","title":"connect(data_stream) async","text":"Connects to the broker running in StreamPipes and creates a subscription.
PARAMETER DESCRIPTIONdata_stream Contains the meta information (resources) for a data stream.
TYPE: DataStream
None"},{"location":"reference/functions/broker/nats/nats_consumer/#streampipes.functions.broker.nats.nats_consumer.NatsConsumer.disconnect","title":"disconnect() async","text":"Closes the connection to the server.
RETURNS DESCRIPTIONNone"},{"location":"reference/functions/broker/nats/nats_consumer/#streampipes.functions.broker.nats.nats_consumer.NatsConsumer.get_message","title":"get_message()","text":"Get the published messages of the subscription.
RETURNS DESCRIPTIONmessage_iterator An async iterator for the messages.
TYPE: AsyncIterator
NatsPublisher","text":" Bases: Publisher
Implementation of a publisher for NATS
"},{"location":"reference/functions/broker/nats/nats_publisher/#streampipes.functions.broker.nats.nats_publisher.NatsPublisher.connect","title":"connect(data_stream) async","text":"Connects to the broker running in StreamPipes.
PARAMETER DESCRIPTIONdata_stream Contains the meta information (resources) for a data stream.
TYPE: DataStream
None"},{"location":"reference/functions/broker/nats/nats_publisher/#streampipes.functions.broker.nats.nats_publisher.NatsPublisher.disconnect","title":"disconnect() async","text":"Closes the connection to the server.
RETURNS DESCRIPTIONNone"},{"location":"reference/functions/broker/nats/nats_publisher/#streampipes.functions.broker.nats.nats_publisher.NatsPublisher.publish_event","title":"publish_event(event) async","text":"Publish an event to a connected data stream.
PARAMETER DESCRIPTIONevent The event to be published.
TYPE: Dict[str, Any]
None"},{"location":"reference/functions/utils/async_iter_handler/","title":"Async iter handler","text":""},{"location":"reference/functions/utils/async_iter_handler/#streampipes.functions.utils.async_iter_handler.AsyncIterHandler","title":"AsyncIterHandler","text":"Handles asynchronous iterators to get every message after another in parallel.
"},{"location":"reference/functions/utils/async_iter_handler/#streampipes.functions.utils.async_iter_handler.AsyncIterHandler.anext","title":"anext(stream_id, message) async staticmethod","text":"Gets the next message from an AsyncIterator.
PARAMETER DESCRIPTIONstream_id The id of the data stream which the message belongs to.
TYPE: str
message An asynchronous iterator that contains the messages.
TYPE: AsyncIterator
result Tuple of the stream id und next message or (\"stop\", None) if no message is left.
TYPE: Tuple[str, Optional[Any]]
combine_async_messages(messages) async staticmethod","text":"Continuously gets the next published message from multiple AsyncIterators in parallel.
PARAMETER DESCRIPTIONmessages A dictionary with an asynchronous iterator for every stream id.
TYPE: Dict[str, AsyncIterator]
message Description of the anonymous integer return value.
TYPE:: Tuple[str, Any]
DataStreamContext(functions, schema, broker)","text":"Container for the context of a data stream.
PARAMETER DESCRIPTIONfunctions StreamPipes Functions which require the data of this data stream.
TYPE: List[StreamPipesFunction]
schema The schema of this data stream.
TYPE: DataStream
broker The consumer to connect to this data stream.
TYPE: Consumer
add_function(function)","text":"Adds a new StreamPipes Function.
PARAMETER DESCRIPTIONfunction StreamPipesFunction which requires this data stream.
TYPE: StreamPipesFunction
None"},{"location":"reference/functions/utils/data_stream_generator/","title":"Data stream generator","text":""},{"location":"reference/functions/utils/data_stream_generator/#streampipes.functions.utils.data_stream_generator.RuntimeType","title":"RuntimeType","text":" Bases: Enum
Runtime type names for the attributes of a data stream.
ATTRIBUTE DESCRIPTIONSTRING
BOOLEAN
DOUBLE
FLOAT
INTEGER
LONG "},{"location":"reference/functions/utils/data_stream_generator/#streampipes.functions.utils.data_stream_generator.create_data_stream","title":"
create_data_stream(name, attributes, stream_id=None, broker=SupportedBroker.NATS)","text":"Creates a data stream
PARAMETER DESCRIPTIONname Name of the data stream to be shown at the UI.
TYPE: str
attributes Name and types of the attributes.
TYPE: Dict[str, str]
stream_id The id of this data stream.
TYPE: Optional[str] DEFAULT: None
data_stream The created data stream
TYPE: DataStream
FunctionContext(function_id, schema, client, streams)","text":"Container for the context of a StreamPipesFunction.
PARAMETER DESCRIPTIONfunction_id The id of this function.
TYPE: str
schema A dictionary which contains the schema of a data stream for each stream id.
TYPE: Dict[str, DataStream]
client The client to interact with the API.
TYPE: StreamPipesClient
streams The ids of the streams needed by this function.
TYPE: List[str]
add_data_stream_schema(stream_id, data_stream)","text":"Adds a new data stream for a new stream id.
PARAMETER DESCRIPTIONstream_id The id of the data stream.
TYPE: str
data_stream The schema of the data stream.
TYPE: DataStream
None"},{"location":"reference/model/common/","title":"Common","text":"Classes of the StreamPipes data model that are commonly shared.
"},{"location":"reference/model/common/#streampipes.model.common.ApplicationLink","title":"ApplicationLink","text":" Bases: BasicModel
Data model of an ApplicationLink in compliance with the StreamPipes Backend.
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/common/#streampipes.model.common.BaseElement","title":"BaseElement","text":" Bases: BasicModel
Structure of a basic element in the StreamPipes Backend.
"},{"location":"reference/model/common/#streampipes.model.common.BaseElement.Config","title":"Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/common/#streampipes.model.common.BasicModel","title":"BasicModel","text":" Bases: BaseModel
Basic model class used for the whole Python StreamPipes data model.
"},{"location":"reference/model/common/#streampipes.model.common.BasicModel.Config","title":"Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/common/#streampipes.model.common.EventGrounding","title":"EventGrounding","text":" Bases: BasicModel
Data model of an EventGrounding in compliance to with StreamPipes Backend.
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/common/#streampipes.model.common.EventProperty","title":"EventProperty","text":" Bases: BasicModel
Data model of an EventProperty in compliance with the StreamPipes Backend.
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/common/#streampipes.model.common.EventSchema","title":"EventSchema","text":" Bases: BasicModel
Data model of an EventSchema in compliance with the StreamPipes Backend.
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/common/#streampipes.model.common.MeasurementCapability","title":"MeasurementCapability","text":" Bases: BasicModel
Data model of a MeasurementCapability in compliance with the StreamPipes Backend.
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/common/#streampipes.model.common.MeasurementObject","title":"MeasurementObject","text":" Bases: BasicModel
Data model of a MeasurementObject in compliance with the StreamPipes Backend.
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/common/#streampipes.model.common.TopicDefinition","title":"TopicDefinition","text":" Bases: BasicModel
Data model of a TopicDefinition in compliance with the StreamPipes Backend.
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/common/#streampipes.model.common.TransportFormat","title":"TransportFormat","text":" Bases: BasicModel
Data model of a TransportFormat in compliance with the StreamPipes Backend.
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/common/#streampipes.model.common.TransportProtocol","title":"TransportProtocol","text":" Bases: BasicModel
Data model of a TransportProtocol in compliance with the StreamPipes Backend.
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/common/#streampipes.model.common.ValueSpecification","title":"ValueSpecification","text":" Bases: BasicModel
Data model of an ValueSpecification in compliance with the StreamPipes Backend.
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/common/#streampipes.model.common.random_letters","title":"random_letters(n)","text":"Generates a string consisting of random letters.
PARAMETER DESCRIPTIONn number of letters
TYPE: int
rand_str String consisting of n random letters
TYPE: str
Implementation of a resource container for the data lake measures endpoint.
"},{"location":"reference/model/container/data_lake_measures/#streampipes.model.container.data_lake_measures.DataLakeMeasures","title":"DataLakeMeasures(resources)","text":" Bases: ResourceContainer
Implementation of the resource container for the data lake measures endpoint.
This resource container is a collection of data lake measures returned by the StreamPipes API. It is capable of parsing the response content directly into a list of queried DataLakeMeasure. Furthermore, the resource container makes them accessible in a pythonic manner.
from_json(json_string) classmethod","text":"Creates a ResourceContainer from the given JSON string.
json_string The JSON string returned from the StreamPipes API.
TYPE: str
container Instance of the container derived from the JSON definition
TYPE: ResourceContainer
StreamPipesDataModelError If a resource cannot be mapped to the corresponding Python data model.
StreamPipesResourceContainerJSONError If JSON response cannot be parsed to a ResourceContainer.
to_dicts(use_source_names=False)","text":"Returns the contained resources as list of dictionaries.
PARAMETER DESCRIPTIONuse_source_names Determines whether the field names are named in Python style (=False) or as originally named by StreamPipes (=True).
TYPE: bool DEFAULT: False
dictionary_list List of resources in dictionary representation. If use_source_names equals True the keys are named as in the StreamPipes backend.
TYPE: List[Dict[str, Any]]
to_json()","text":"Returns the resource container in the StreamPipes JSON representation.
RETURNS DESCRIPTION json_string: str JSON representation of the resource container where key names are equal to keys used in the StreamPipes backend
"},{"location":"reference/model/container/data_lake_measures/#streampipes.model.container.data_lake_measures.DataLakeMeasures.to_pandas","title":"to_pandas()","text":"Returns the resource container in representation of a Pandas Dataframe.
RETURNS DESCRIPTIONresource_container_df Representation of the resource container as pandas DataFrame
TYPE: DataFrame
Implementation of a resource container for the data streams endpoint.
"},{"location":"reference/model/container/data_streams/#streampipes.model.container.data_streams.DataStreams","title":"DataStreams(resources)","text":" Bases: ResourceContainer
Implementation of the resource container for the data stream endpoint.
This resource container is a collection of data streams returned by the StreamPipes API. It is capable of parsing the response content directly into a list of queried DataStream. Furthermore, the resource container makes them accessible in a pythonic manner.
from_json(json_string) classmethod","text":"Creates a ResourceContainer from the given JSON string.
json_string The JSON string returned from the StreamPipes API.
TYPE: str
container Instance of the container derived from the JSON definition
TYPE: ResourceContainer
StreamPipesDataModelError If a resource cannot be mapped to the corresponding Python data model.
StreamPipesResourceContainerJSONError If JSON response cannot be parsed to a ResourceContainer.
to_dicts(use_source_names=False)","text":"Returns the contained resources as list of dictionaries.
PARAMETER DESCRIPTIONuse_source_names Determines whether the field names are named in Python style (=False) or as originally named by StreamPipes (=True).
TYPE: bool DEFAULT: False
dictionary_list List of resources in dictionary representation. If use_source_names equals True the keys are named as in the StreamPipes backend.
TYPE: List[Dict[str, Any]]
to_json()","text":"Returns the resource container in the StreamPipes JSON representation.
RETURNS DESCRIPTION json_string: str JSON representation of the resource container where key names are equal to keys used in the StreamPipes backend
"},{"location":"reference/model/container/data_streams/#streampipes.model.container.data_streams.DataStreams.to_pandas","title":"to_pandas()","text":"Returns the resource container in representation of a Pandas Dataframe.
RETURNS DESCRIPTIONresource_container_df Representation of the resource container as pandas DataFrame
TYPE: DataFrame
General and abstract implementation for a resource container.
A resource container is a collection of resources returned by the StreamPipes API. It is capable of parsing the response content directly into a list of queried resources. Furthermore, the resource container makes them accessible in a pythonic manner.
"},{"location":"reference/model/container/resource_container/#streampipes.model.container.resource_container.ResourceContainer","title":"ResourceContainer(resources)","text":" Bases: ABC
General and abstract implementation for a resource container.
A resource container is a collection of resources returned by the StreamPipes API. It is capable of parsing the response content directly into a list of queried resources. Furthermore, the resource container makes them accessible in a pythonic manner.
PARAMETER DESCRIPTIONresources A list of resources to be contained in the ResourceContainer.
TYPE: List[Resource]
from_json(json_string) classmethod","text":"Creates a ResourceContainer from the given JSON string.
json_string The JSON string returned from the StreamPipes API.
TYPE: str
container Instance of the container derived from the JSON definition
TYPE: ResourceContainer
StreamPipesDataModelError If a resource cannot be mapped to the corresponding Python data model.
StreamPipesResourceContainerJSONError If JSON response cannot be parsed to a ResourceContainer.
to_dicts(use_source_names=False)","text":"Returns the contained resources as list of dictionaries.
PARAMETER DESCRIPTIONuse_source_names Determines whether the field names are named in Python style (=False) or as originally named by StreamPipes (=True).
TYPE: bool DEFAULT: False
dictionary_list List of resources in dictionary representation. If use_source_names equals True the keys are named as in the StreamPipes backend.
TYPE: List[Dict[str, Any]]
to_json()","text":"Returns the resource container in the StreamPipes JSON representation.
RETURNS DESCRIPTION json_string: str JSON representation of the resource container where key names are equal to keys used in the StreamPipes backend
"},{"location":"reference/model/container/resource_container/#streampipes.model.container.resource_container.ResourceContainer.to_pandas","title":"to_pandas()","text":"Returns the resource container in representation of a Pandas Dataframe.
RETURNS DESCRIPTIONresource_container_df Representation of the resource container as pandas DataFrame
TYPE: DataFrame
StreamPipesDataModelError(validation_error)","text":" Bases: Exception
A custom exception to be raised when a validation error occurs during the parsing of StreamPipes API responses.
PARAMETER DESCRIPTIONvalidation_error The validation error thrown by Pydantic during parsing.
TYPE: ValidationError
StreamPipesResourceContainerJSONError(container_name, json_string)","text":" Bases: Exception
A custom exception to be raised when the returned JSON string does not suit to the structure of resource container.
PARAMETER DESCRIPTIONcontainer_name The class name of the resource container where the invalid data structure was detected.
TYPE: str
json_string The JSON string that has been tried to parse.
TYPE: str
Implementation of a resource container for the versions endpoint.
"},{"location":"reference/model/container/versions/#streampipes.model.container.versions.Versions","title":"Versions(resources)","text":" Bases: ResourceContainer
Implementation of the resource container for the versions endpoint.
This resource container is a collection of versions returned by the StreamPipes API. It is capable of parsing the response content directly into a list of queried Version. Furthermore, the resource container makes them accessible in a pythonic manner.
resources A list of resources (Version) to be contained in the ResourceContainer.
TYPE: List[Resource]
from_json(json_string) classmethod","text":"Creates a ResourceContainer from the given JSON string.
json_string The JSON string returned from the StreamPipes API.
TYPE: str
container Instance of the container derived from the JSON definition
TYPE: ResourceContainer
StreamPipesDataModelError If a resource cannot be mapped to the corresponding Python data model.
StreamPipesResourceContainerJSONError If JSON response cannot be parsed to a ResourceContainer.
to_dicts(use_source_names=False)","text":"Returns the contained resources as list of dictionaries.
PARAMETER DESCRIPTIONuse_source_names Determines whether the field names are named in Python style (=False) or as originally named by StreamPipes (=True).
TYPE: bool DEFAULT: False
dictionary_list List of resources in dictionary representation. If use_source_names equals True the keys are named as in the StreamPipes backend.
TYPE: List[Dict[str, Any]]
to_json()","text":"Returns the resource container in the StreamPipes JSON representation.
RETURNS DESCRIPTION json_string: str JSON representation of the resource container where key names are equal to keys used in the StreamPipes backend
"},{"location":"reference/model/container/versions/#streampipes.model.container.versions.Versions.to_pandas","title":"to_pandas()","text":"Returns the resource container in representation of a Pandas Dataframe.
RETURNS DESCRIPTIONresource_container_df Representation of the resource container as pandas DataFrame
TYPE: DataFrame
DataLakeMeasure","text":" Bases: Resource
Implementation of a resource for data lake measures.
This resource defines the data model used by resource container (model.container.DataLakeMeasures). It inherits from Pydantic's BaseModel to get all its superpowers, which are used to parse, validate the API response, and to easily switch between the Python representation (both serialized and deserialized) and Java representation (serialized only).
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/resource/data_lake_measure/#streampipes.model.resource.data_lake_measure.DataLakeMeasure.convert_to_pandas_representation","title":"convert_to_pandas_representation()","text":"Returns the dictionary representation of a data lake measure to be used when creating a pandas Dataframe.
It excludes the following fields: element_id, event_schema, schema_version. Instead of the whole event schema the number of event properties contained is returned with the column name num_event_properties.
pandas_repr Pandas representation of the resource as a dictionary, which is then used by the respource container to create a data frame from a collection of resources.
TYPE: Dict[str, Any]
to_dict(use_source_names=True)","text":"Returns the resource in dictionary representation.
PARAMETER DESCRIPTIONuse_source_names Indicates if the dictionary keys are in python representation or equally named to the StreamPipes backend
DEFAULT: True
resource The resource as dictionary representation
TYPE: Dict[str, Any]
DataSeries","text":" Bases: Resource
Implementation of a resource for data series. This resource defines the data model used by its resource container(model.container.DataLakeMeasures). It inherits from Pydantic's BaseModel to get all its superpowers, which are used to parse, validate the API response and to easily switch between the Python representation (both serialized and deserialized) and Java representation (serialized only).
This class will only exist temporarily in it its current appearance since\nthere are some inconsistencies in the StreamPipes API.\nConfig","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/resource/data_series/#streampipes.model.resource.data_series.DataSeries.convert_to_pandas_representation","title":"convert_to_pandas_representation()","text":"Returns the dictionary representation of a data lake series to be used when creating a pandas Dataframe.
It contains only the \"header rows\" (the column names) and \"rows\" that contain the actual data.
RETURNS DESCRIPTIONpandas_repr Dictionary with the keys headers and rows
TYPE: dict[str, Any]
from_json(json_string) classmethod","text":"Creates an instance of DataSeries from a given JSON string.
This method is used by the resource container to parse the JSON response of the StreamPipes API. Currently, it only supports data lake series that consist of exactly one series of data.
PARAMETER DESCRIPTIONjson_string The JSON string the data lake series should be created on.
TYPE: str
DataSeries Instance of DataSeries that is created based on the given JSON string.
StreamPipesUnsupportedDataLakeSeries If the data lake series returned by the StreamPipes API cannot be parsed with the current version of the Python client.
"},{"location":"reference/model/resource/data_series/#streampipes.model.resource.data_series.DataSeries.to_dict","title":"to_dict(use_source_names=True)","text":"Returns the resource in dictionary representation.
PARAMETER DESCRIPTIONuse_source_names Indicates if the dictionary keys are in python representation or equally named to the StreamPipes backend
DEFAULT: True
resource The resource as dictionary representation
TYPE: Dict[str, Any]
to_pandas()","text":"Returns the data lake series in representation of a Pandas Dataframe.
RETURNS DESCRIPTIONpd The data lake series in form of a pandas dataframe
TYPE: DataFrame
DataStream(**kwargs)","text":" Bases: Resource
Implementation of a resource for data streams.
This resource defines the data model used by resource container (model.container.DataStreams). It inherits from Pydantic's BaseModel to get all its superpowers, which are used to parse, validate the API response and to easily switch between the Python representation (both serialized and deserialized) and Java representation (serialized only).
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/resource/data_stream/#streampipes.model.resource.data_stream.DataStream.convert_to_pandas_representation","title":"convert_to_pandas_representation()","text":"Returns the dictionary representation of a data stream to be used when creating a pandas Dataframe.
RETURNS DESCRIPTIONpandas_repr Pandas representation of the resource as a dictionary, which is then used by the respource container to create a data frame from a collection of resources.
TYPE: Dict[str, Any]
to_dict(use_source_names=True)","text":"Returns the resource in dictionary representation.
PARAMETER DESCRIPTIONuse_source_names Indicates if the dictionary keys are in python representation or equally named to the StreamPipes backend
DEFAULT: True
resource The resource as dictionary representation
TYPE: Dict[str, Any]
StreamPipesUnsupportedDataSeries(reason=None)","text":" Bases: Exception
Exception to be raised when the returned data lake series cannot be parsed with the current implementation of the resource.
"},{"location":"reference/model/resource/function_definition/","title":"Function definition","text":""},{"location":"reference/model/resource/function_definition/#streampipes.model.resource.function_definition.FunctionDefinition","title":"FunctionDefinition","text":" Bases: Resource
Configuration for a StreamPipes Function.
This class maps to the FunctionDefinition class in the StreamPipes model. It contains all metadata that are required to register a function at the StreamPipes backend.
consumed_streams List of data streams the function is consuming from
TYPE: List[str]
function_id identifier object of a StreamPipes function
TYPE: FunctionId
output_data_streams Map off all output data streams added to the function definition
TYPE: Dict[str, DataStream]
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/resource/function_definition/#streampipes.model.resource.function_definition.FunctionDefinition.add_output_data_stream","title":"add_output_data_stream(data_stream)","text":"Adds an output data stream to the function which makes it possible to write data back to StreamPipes.
PARAMETER DESCRIPTIONdata_stream The schema of the output data stream.
TYPE: DataStream
self Instance of the function definition that is extended by the provided DataStream
TYPE: FunctionDefinition
convert_to_pandas_representation()","text":"Returns the dictionary representation of a function definition to be used when creating a pandas Dataframe.
RETURNS DESCRIPTIONpandas_repr Pandas representation of the resource as a dictionary, which is then used by the respource container to create a data frame from a collection of resources.
TYPE: Dict[str, Any]
get_output_data_streams()","text":"Get the output data streams of the function.
RETURNS DESCRIPTIONoutput_streams Dictionary with every known stream id and the related output stream.
TYPE: Dict[str, DataStream]
get_output_stream_ids()","text":"Get the stream ids of the output data streams.
RETURNS DESCRIPTIONoutput_stream_ids List of all stream ids
TYPE: List[str]
to_dict(use_source_names=True)","text":"Returns the resource in dictionary representation.
PARAMETER DESCRIPTIONuse_source_names Indicates if the dictionary keys are in python representation or equally named to the StreamPipes backend
DEFAULT: True
resource The resource as dictionary representation
TYPE: Dict[str, Any]
FunctionId","text":" Bases: BasicModel
Identification object for a StreamPipes function.
Maps to the FunctionId class defined in the StreamPipes model.
id unique identifier of the function instance
TYPE: str
version version of the corresponding function
TYPE: int
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/resource/query_result/","title":"Query result","text":""},{"location":"reference/model/resource/query_result/#streampipes.model.resource.query_result.QueryResult","title":"QueryResult","text":" Bases: Resource
Implementation of a resource for query result. This resource defines the data model used by its resource container(model.container.DataLakeMeasures). It inherits from Pydantic's BaseModel to get all its superpowers, which are used to parse, validate the API response and to easily switch between the Python representation (both serialized and deserialized) and Java representation (serialized only).
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/resource/query_result/#streampipes.model.resource.query_result.QueryResult.convert_to_pandas_representation","title":"convert_to_pandas_representation()","text":"Returns the dictionary representation of a data lake series to be used when creating a pandas Dataframe.
It contains only the \"header rows\" (the column names) and \"rows\" that contain the actual data.
RETURNS DESCRIPTIONdict Dictionary with the keys headers and rows
StreamPipesUnsupportedDataLakeSeries If the query result returned by the StreamPipes API cannot be converted to the pandas representation
"},{"location":"reference/model/resource/query_result/#streampipes.model.resource.query_result.QueryResult.to_dict","title":"to_dict(use_source_names=True)","text":"Returns the resource in dictionary representation.
PARAMETER DESCRIPTIONuse_source_names Indicates if the dictionary keys are in python representation or equally named to the StreamPipes backend
DEFAULT: True
resource The resource as dictionary representation
TYPE: Dict[str, Any]
to_pandas()","text":"Returns the data lake series in representation of a Pandas Dataframe.
RETURNS DESCRIPTIONdf Pandas df containing the query result
TYPE: DataFrame
General and abstract implementation for a resource.
A resource defines the data model that is used by a resource container (model.container.resourceContainer).
Resource","text":" Bases: ABC, BasicModel
General and abstract implementation for a resource.
A resource defines the data model used by a resource container (model.container.resourceContainer). It inherits from Pydantic's BaseModel to get all its superpowers, which are used to parse, validate the API response and to easily switch between the Python representation (both serialized and deserialized) and Java representation (serialized only).
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/resource/resource/#streampipes.model.resource.resource.Resource.convert_to_pandas_representation","title":"convert_to_pandas_representation() abstractmethod","text":"Returns a dictionary representation to be used when creating a pandas Dataframe.
RETURNS DESCRIPTIONpandas_repr Pandas representation of the resource as a dictionary, which is then used by the respource container to create a data frame from a collection of resources.
TYPE: Dict[str, Any]
to_dict(use_source_names=True)","text":"Returns the resource in dictionary representation.
PARAMETER DESCRIPTIONuse_source_names Indicates if the dictionary keys are in python representation or equally named to the StreamPipes backend
DEFAULT: True
resource The resource as dictionary representation
TYPE: Dict[str, Any]
Version","text":" Bases: Resource
Metadata about the version of the connected StreamPipes server.
ATTRIBUTE DESCRIPTIONbackend_version version of the StreamPipes backend the client is connected to
TYPE: str
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/resource/version/#streampipes.model.resource.version.Version.convert_to_pandas_representation","title":"convert_to_pandas_representation()","text":"Returns the dictionary representation of the version metadata to be used when creating a pandas Dataframe.
"},{"location":"reference/model/resource/version/#streampipes.model.resource.version.Version.to_dict","title":"to_dict(use_source_names=True)","text":"Returns the resource in dictionary representation.
PARAMETER DESCRIPTIONuse_source_names Indicates if the dictionary keys are in python representation or equally named to the StreamPipes backend
DEFAULT: True
resource The resource as dictionary representation
TYPE: Dict[str, Any]
validate_backend_version(backend_version)","text":"Validates the backend version of the StreamPipes. Sets 'development' if none is returned since this the behavior of StreamPipes backend running in development mode.
"},{"location":"tutorials/1-introduction-to-streampipes-python-client/","title":"Introduction to StreamPipes Python","text":"In\u00a0[\u00a0]: Copied!%pip install streampipes\n%pip install streampipes
If you want to have the current development state you can also execute:
In\u00a0[\u00a0]: Copied!%pip install git+https://github.com/apache/streampipes.git#subdirectory=streampipes-client-python\n%pip install git+https://github.com/apache/streampipes.git#subdirectory=streampipes-client-python
The corresponding documentation can be found here.
In\u00a0[\u00a0]: Copied!from streampipes.client import StreamPipesClient\nfrom streampipes.client.config import StreamPipesClientConfig\nfrom streampipes.client.credential_provider import StreamPipesApiKeyCredentials\nfrom streampipes.client import StreamPipesClient from streampipes.client.config import StreamPipesClientConfig from streampipes.client.credential_provider import StreamPipesApiKeyCredentials In\u00a0[\u00a0]: Copied!
config = StreamPipesClientConfig(\n credential_provider=StreamPipesApiKeyCredentials(\n username=\"test@streampipes.apache.org\",\n api_key=\"API-KEY\",\n ),\n host_address=\"localhost\",\n https_disabled=True,\n port=80\n)\nconfig = StreamPipesClientConfig( credential_provider=StreamPipesApiKeyCredentials( username=\"test@streampipes.apache.org\", api_key=\"API-KEY\", ), host_address=\"localhost\", https_disabled=True, port=80 )
Please be aware that connecting to StreamPipes via a https connection is currently not supported by the Python client.
Providing secrets like the api_key as plaintext in the source code is an anti-pattern. This is why the StreamPipes client also supports passing the required secrets as environment variables. To do so, you must initialize the credential provider like the following:
StreamPipesApiKeyCredentials()\nStreamPipesApiKeyCredentials()
To ensure that the above code works, you must set the environment variables as expected. This can be done as follows:
In\u00a0[\u00a0]: Copied!import os\nos.environ[\"SP_USERNAME\"] = \"admin@streampipes.apache.org\"\nos.environ[\"SP_API_KEY\"] = \"XXX\"\nimport os os.environ[\"SP_USERNAME\"] = \"admin@streampipes.apache.org\" os.environ[\"SP_API_KEY\"] = \"XXX\"
Having the config ready, we can now initialize the actual client.
client = StreamPipesClient(client_config=config)\nclient = StreamPipesClient(client_config=config)
That's already it. You can check if everything works out by using the following command:
In\u00a0[6]: Copied!client.describe()\nclient.describe()
2023-02-24 17:05:49,398 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:167] [_make_request] - Successfully retrieved all resources.\n2023-02-24 17:05:49,457 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:167] [_make_request] - Successfully retrieved all resources.\n\nHi there!\nYou are connected to a StreamPipes instance running at http://localhost:80.\nThe following StreamPipes resources are available with this client:\n1x DataLakeMeasures\n1x DataStreams\n
This prints you a short textual description of the connected StreamPipes instance to the console.
The created client instance serves as the central point of interaction with StreamPipes. You can invoke a variety of commands directly on this object.
Are you curious now how you actually can get data out of StreamPipes and make use of it with Python? Then check out the next tutorial on extracting Data from the StreamPipes data lake.
Thanks for reading this introductory tutorial. We hope you like it and would love to receive some feedback from you. Just go to our GitHub discussion page and let us know your impression. We'll read and react to them all, we promise!
"},{"location":"tutorials/1-introduction-to-streampipes-python-client/#Introduction-to-StreamPipes-Python","title":"Introduction to StreamPipes Python\u00b6","text":""},{"location":"tutorials/1-introduction-to-streampipes-python-client/#Why-there-is-an-extra-Python-library-for-StreamPipes?","title":"Why there is an extra Python library for StreamPipes?\u00b6","text":"Apache StreamPipes aims to enable non-technical users to connect and analyze IoT data streams. To achieve this, it provides an easy-to-use and convenient user interface that allows one to connect to an IoT data source and create some visual graphs within a few minutes. While this is the primary use case for Apache StreamPipes, it also offers significant value to those interested in data analysis or data science with IoT data, without the need to handle the complexities of extracting data from devices in a suitable format. In this scenario, StreamPipes helps you connect to your data source and extract the data for you. You then can make the data available outside StreamPipes by writing it into an external source, such as a database, Kafka, etc. While this requires another component, you can also extract your data directly from StreamPipes programmatically using the StreamPipes API. For convenience, we also provide you with a StreamPipes client both available for Java and Python. Specifically with StreamPipes Python, we want to address the amazing data analytics and data science community in Python and benefit from the great universe of Python libraries out there.
"},{"location":"tutorials/1-introduction-to-streampipes-python-client/#How-to-install-StreamPipes-Python?","title":"How to install StreamPipes Python?\u00b6","text":"Simply use the following pip command:
In case you want to reproduce the first two tutorials exactly on your end, you need to create a simple pipeline in StreamPipes like demonstrated below.
"},{"location":"tutorials/1-introduction-to-streampipes-python-client/#How-to-configure-the-Python-client","title":"How to configure the Python client\u00b6","text":"In order to access the resources available in StreamPipes, one must be able to authenticate against the backend. For this purpose, the client so far only supports the authentication via an API token that can be generated via the StreamPipes UI, as you can see below.
Having generated the API token, one can directly start initializing a client instance as follows:
"},{"location":"tutorials/2-extracting-data-from-the-streampipes-data-lake/","title":"Extracting Data from the StreamPipes data lake","text":"In\u00a0[1]: Copied!from streampipes.client import StreamPipesClient\nfrom streampipes.client.config import StreamPipesClientConfig\nfrom streampipes.client.credential_provider import StreamPipesApiKeyCredentials\nfrom streampipes.client import StreamPipesClient from streampipes.client.config import StreamPipesClientConfig from streampipes.client.credential_provider import StreamPipesApiKeyCredentials In\u00a0[\u00a0]: Copied!
# if you want all necessary dependencies required for this tutorial to be installed,\n# you can simply execute the following command\n%pip install matplotlib streampipes\n# if you want all necessary dependencies required for this tutorial to be installed, # you can simply execute the following command %pip install matplotlib streampipes In\u00a0[2]: Copied!
import os\nos.environ[\"SP_USERNAME\"] = \"admin@streampipes.apache.org\"\nos.environ[\"SP_API_KEY\"] = \"XXX\"\nimport os os.environ[\"SP_USERNAME\"] = \"admin@streampipes.apache.org\" os.environ[\"SP_API_KEY\"] = \"XXX\" In\u00a0[3]: Copied!
config = StreamPipesClientConfig(\n credential_provider=StreamPipesApiKeyCredentials(),\n host_address=\"localhost\",\n https_disabled=True,\n port=80\n)\nconfig = StreamPipesClientConfig( credential_provider=StreamPipesApiKeyCredentials(), host_address=\"localhost\", https_disabled=True, port=80 ) In\u00a0[4]: Copied!
client = StreamPipesClient(client_config=config)\nclient = StreamPipesClient(client_config=config)
2023-02-24 17:34:25,860 - streampipes.client.client - [INFO] - [client.py:128] [_set_up_logging] - Logging successfully initialized with logging level INFO.\n
As a first step, we want to get an overview about all data available in the data lake. The data is stored as so-called measures, which refer to a data stream stored in the data lake. For his purpose we use the all() method of the dataLakeMeasure endpoint.
data_lake_measures = client.dataLakeMeasureApi.all()\ndata_lake_measures = client.dataLakeMeasureApi.all()
2023-02-24 17:34:25,929 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:167] [_make_request] - Successfully retrieved all resources.\n
So let's see how many measures are available:
In\u00a0[6]: Copied!len(data_lake_measures)\nlen(data_lake_measures) Out[6]:
2
All resources of the StreamPipes Python client support the standard Python expressions. If not, please let us know.
In\u00a0[7]: Copied!data_lake_measures[-1]\ndata_lake_measures[-1] Out[7]:
DataLakeMeasure(element_id='3cb6b5e6f107452483d1fd2ccf4bf9f9', measure_name='test', timestamp_field='s0::timestamp', event_schema=EventSchema(event_properties=[EventProperty(class_name='org.apache.streampipes.model.schema.EventPropertyPrimitive', element_id='sp:eventproperty:EiFnkL', label='Density', description='Denotes the current density of the fluid', runtime_name='density', required=False, domain_properties=['http://schema.org/Number'], property_scope='MEASUREMENT_PROPERTY', index=5, runtime_id=None, runtime_type='http://www.w3.org/2001/XMLSchema#float', measurement_unit=None, value_specification=None), EventProperty(class_name='org.apache.streampipes.model.schema.EventPropertyPrimitive', element_id='sp:eventproperty:ghSkQI', label='Mass Flow', description='Denotes the current mass flow in the sensor', runtime_name='mass_flow', required=False, domain_properties=['http://schema.org/Number'], property_scope='MEASUREMENT_PROPERTY', index=2, runtime_id=None, runtime_type='http://www.w3.org/2001/XMLSchema#float', measurement_unit=None, value_specification=None), EventProperty(class_name='org.apache.streampipes.model.schema.EventPropertyPrimitive', element_id='sp:eventproperty:cQAUry', label='Sensor ID', description='The ID of the sensor', runtime_name='sensorId', required=False, domain_properties=['https://streampipes.org/vocabulary/examples/watertank/v1/hasSensorId'], property_scope='DIMENSION_PROPERTY', index=1, runtime_id=None, runtime_type='http://www.w3.org/2001/XMLSchema#string', measurement_unit=None, value_specification=None), EventProperty(class_name='org.apache.streampipes.model.schema.EventPropertyPrimitive', element_id='sp:eventproperty:pbPMyL', label='Sensor Fault Flags', description='Any fault flags of the sensors', runtime_name='sensor_fault_flags', required=False, domain_properties=['http://schema.org/Boolean'], property_scope='MEASUREMENT_PROPERTY', index=6, runtime_id=None, runtime_type='http://www.w3.org/2001/XMLSchema#boolean', measurement_unit=None, value_specification=None), EventProperty(class_name='org.apache.streampipes.model.schema.EventPropertyPrimitive', element_id='sp:eventproperty:Qmayhw', label='Temperature', description='Denotes the current temperature in degrees celsius', runtime_name='temperature', required=False, domain_properties=['http://schema.org/Number'], property_scope='MEASUREMENT_PROPERTY', index=4, runtime_id=None, runtime_type='http://www.w3.org/2001/XMLSchema#float', measurement_unit='http://qudt.org/vocab/unit#DegreeCelsius', value_specification=ValueSpecification(class_name='org.apache.streampipes.model.schema.QuantitativeValue', element_id=None, min_value=0, max_value=100, step=0.1)), EventProperty(class_name='org.apache.streampipes.model.schema.EventPropertyPrimitive', element_id='sp:eventproperty:YQYhjd', label='Volume Flow', description='Denotes the current volume flow', runtime_name='volume_flow', required=False, domain_properties=['http://schema.org/Number'], property_scope='MEASUREMENT_PROPERTY', index=3, runtime_id=None, runtime_type='http://www.w3.org/2001/XMLSchema#float', measurement_unit=None, value_specification=None)]), pipeline_id=None, pipeline_name=None, pipeline_is_running=False, schema_version='1.1')
To get a more comprehensive overview, you can take a look at the pandas representation:
display(data_lake_measures.to_pandas())\ndisplay(data_lake_measures.to_pandas()) measure_name timestamp_field pipeline_id pipeline_name pipeline_is_running num_event_properties 0 flow-rate s0::timestamp None None False 6 1 test s0::timestamp None None False 6
So far, we have only retrieved metadata about the available data lake measure. In the following, we will access the actual data of the measure flow-rate.
For this purpose, we will use the get() method of the dataLakeMeasure endpoint.
flow_rate_measure = client.dataLakeMeasureApi.get(identifier=\"flow-rate\")\nflow_rate_measure = client.dataLakeMeasureApi.get(identifier=\"flow-rate\")
2023-02-24 17:34:26,020 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:167] [_make_request] - Successfully retrieved all resources.\n
For further processing, the easiest way is to turn the data measure into a pandas DataFrame.
flow_rate_pd = flow_rate_measure.to_pandas()\nflow_rate_pd = flow_rate_measure.to_pandas()
Let's see how many data points we got...
In\u00a0[11]: Copied!len(flow_rate_pd)\nlen(flow_rate_pd) Out[11]:
1000
... and get a first overview
In\u00a0[12]: Copied!flow_rate_pd.describe()\nflow_rate_pd.describe() Out[12]: density mass_flow temperature volume_flow count 1000.000000 1000.000000 1000.000000 1000.000000 mean 45.560337 5.457014 45.480231 5.659558 std 3.201544 3.184959 3.132878 3.122437 min 40.007698 0.004867 40.000992 0.039422 25% 42.819497 2.654101 42.754623 3.021625 50% 45.679264 5.382355 45.435944 5.572553 75% 48.206881 8.183144 48.248473 8.338209 max 50.998310 10.986015 50.964909 10.998676
As a final step, we want to create a plot of both attributes.
In\u00a0[13]: Copied!import matplotlib.pyplot as plt\nflow_rate_pd.plot(y=[\"mass_flow\", \"temperature\"])\nplt.show()\nimport matplotlib.pyplot as plt flow_rate_pd.plot(y=[\"mass_flow\", \"temperature\"]) plt.show()
For data lake measurements, the get() method is even more powerful than simply returning all the data for a given data lake measurement. We will look at a selection of these below. The full list of supported parameters can be found in the docs. Let's start by referring to the graph we created above, where we use only two columns of our data lake measurement. If we already know this, we can directly restrict the queried data to a subset of columns by using the columns parameter. columns takes a list of column names as a comma-separated string:
flow_rate_pd = client.dataLakeMeasureApi.get(identifier=\"flow-rate\", columns=\"mass_flow,temperature\").to_pandas()\nflow_rate_pd\nflow_rate_pd = client.dataLakeMeasureApi.get(identifier=\"flow-rate\", columns=\"mass_flow,temperature\").to_pandas() flow_rate_pd
2023-02-24 17:34:26,492 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:167] [_make_request] - Successfully retrieved all resources.\nOut[14]: timestamp mass_flow temperature 0 2023-02-24T16:19:41.472Z 3.309556 44.448483 1 2023-02-24T16:19:41.482Z 5.608580 40.322033 2 2023-02-24T16:19:41.493Z 7.692881 49.239639 3 2023-02-24T16:19:41.503Z 3.632898 49.933754 4 2023-02-24T16:19:41.513Z 0.711260 50.106617 ... ... ... ... 995 2023-02-24T16:19:52.927Z 1.740114 46.558231 996 2023-02-24T16:19:52.94Z 7.211723 48.048622 997 2023-02-24T16:19:52.952Z 7.770180 48.188026 998 2023-02-24T16:19:52.965Z 4.458602 48.280899 999 2023-02-24T16:19:52.977Z 2.592060 47.505951
1000 rows \u00d7 3 columns
By default, the client returns only the first one thousand records of a Data Lake measurement. This can be changed by passing a concrete value for the limit parameter:
flow_rate_pd = client.dataLakeMeasureApi.get(identifier=\"flow-rate\", limit=10000).to_pandas()\nlen(flow_rate_pd)\nflow_rate_pd = client.dataLakeMeasureApi.get(identifier=\"flow-rate\", limit=10000).to_pandas() len(flow_rate_pd)
2023-02-24 17:34:26,736 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:167] [_make_request] - Successfully retrieved all resources.\nOut[15]:
9528
If you want your data to be selected by time of occurrence rather than quantity, you can specify your time window by passing the start_date and end_date parameters:
from datetime import datetime\nflow_rate_pd = client.dataLakeMeasureApi.get(\n identifier=\"flow-rate\",\n start_date=datetime(year=2023, month=2, day=24, hour=17, minute=21, second=0),\n end_date=datetime(year=2023, month=2, day=24, hour=17, minute=21, second=1),\n ).to_pandas()\nflow_rate_pd.plot(y=[\"mass_flow\", \"temperature\"])\nplt.show()\nfrom datetime import datetime flow_rate_pd = client.dataLakeMeasureApi.get( identifier=\"flow-rate\", start_date=datetime(year=2023, month=2, day=24, hour=17, minute=21, second=0), end_date=datetime(year=2023, month=2, day=24, hour=17, minute=21, second=1), ).to_pandas() flow_rate_pd.plot(y=[\"mass_flow\", \"temperature\"]) plt.show()
2023-02-24 17:34:26,899 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:167] [_make_request] - Successfully retrieved all resources.\n
... from this point on, we leave all future processing of the data up to your creativity. Keep in mind: the general syntax used in this tutorial (all(), to_pandas(), get()) applies to all endpoints and associated resources of the StreamPipes Python client.
If you get further and create exciting stuff with data extracted from StreamPipes please let us know. We are thrilled to see what you as a community will build with the provided client. Furthermore, don't hesitate to discuss feature requests to extend the current functionality with us.
For now, that's all about the StreamPipes client. Read the next tutorial (Getting live data from the StreamPipes data stream) if you are interested in making use of the powerful StreamPipes functions to interact with StreamPipes in an event-based manner.
How do you like this tutorial? We hope you like it and would love to receive some feedback from you. Just go to our GitHub discussion page and let us know your impression. We'll read and react to them all, we promise!
"},{"location":"tutorials/2-extracting-data-from-the-streampipes-data-lake/#Extracting-Data-from-the-StreamPipes-data-lake","title":"Extracting Data from the StreamPipes data lake\u00b6","text":"In the first tutorial (Introduction to the StreamPipes Python client) we took the first steps with the StreamPipes Python client and learned how to set everything up. Now we are ready to get started and want to retrieve some data out of StreamPipes. In this tutorial, we'll focus on the StreamPipes Data Lake, the component where StreamPipes stores data internally. To get started, we'll use the client instance created in the first tutorial.
Note As of now we mainly developed the support for StreamPipes functions using NATS as messaging protocol. Consequently, this setup is tested most and should work flawlessly. Visit our first-steps page to see how to start StreamPipes accordingly. Anyhow, you can also use the other brokers that are currently supported in StreamPipes Python. In case you observe any problems, please reach out to us and file us an issue on GitHub.
In\u00a0[1]: Copied!from streampipes.client import StreamPipesClient\nfrom streampipes.client.config import StreamPipesClientConfig\nfrom streampipes.client.credential_provider import StreamPipesApiKeyCredentials\nfrom streampipes.client import StreamPipesClient from streampipes.client.config import StreamPipesClientConfig from streampipes.client.credential_provider import StreamPipesApiKeyCredentials In\u00a0[\u00a0]: Copied!
# You can install all required libraries for this tutorial with the following command\n%pip install matplotlib ipython streampipes\n# You can install all required libraries for this tutorial with the following command %pip install matplotlib ipython streampipes In\u00a0[2]: Copied!
import os\n\nos.environ[\"SP_USERNAME\"] = \"admin@streampipes.apache.org\"\nos.environ[\"SP_API_KEY\"] = \"XXX\"\n\n# Use this if you work locally:\nos.environ[\"BROKER-HOST\"] = \"localhost\" \nos.environ[\"KAFKA-PORT\"] = \"9094\" # When using Kafka as message broker. If Kafka is not running on localhost, KAFKA_ADVERTISED_LISTENERS should be adjusted to the external address\nimport os os.environ[\"SP_USERNAME\"] = \"admin@streampipes.apache.org\" os.environ[\"SP_API_KEY\"] = \"XXX\" # Use this if you work locally: os.environ[\"BROKER-HOST\"] = \"localhost\" os.environ[\"KAFKA-PORT\"] = \"9094\" # When using Kafka as message broker. If Kafka is not running on localhost, KAFKA_ADVERTISED_LISTENERS should be adjusted to the external address In\u00a0[3]: Copied!
client_config = StreamPipesClientConfig(\n credential_provider=StreamPipesApiKeyCredentials(),\n host_address=\"localhost\",\n port=80,\n https_disabled=True,\n)\nclient = StreamPipesClient(client_config=client_config)\nclient_config = StreamPipesClientConfig( credential_provider=StreamPipesApiKeyCredentials(), host_address=\"localhost\", port=80, https_disabled=True, ) client = StreamPipesClient(client_config=client_config)
2022-12-14 10:43:37,664 - streampipes.client.client - [INFO] - [client.py:127] [_set_up_logging] - Logging successfully initialized with logging level INFO.\n
Now we can have a look at the available data streams. We can choose one or more stream to receive the data from and copy their element_id.
client.dataStreamApi.all().to_pandas()\nclient.dataStreamApi.all().to_pandas()
2022-12-14 10:43:39,944 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:153] [_make_request] - Successfully retrieved all resources.\nOut[4]: element_id name description icon_url app_id includes_assets includes_locales internally_managed measurement_object index ... uri dom num_transport_protocols num_measurement_capability num_application_links num_included_assets num_connected_to num_category num_event_properties num_included_locales 0 urn:streampipes.apache.org:eventstream:HHoidJ Test2 None None False False True None 0 ... urn:streampipes.apache.org:eventstream:HHoidJ None 1 0 0 0 0 0 7 0 1 urn:streampipes.apache.org:eventstream:uPDKLI Test None None False False True None 0 ... urn:streampipes.apache.org:eventstream:uPDKLI None 1 0 0 0 0 0 7 0
2 rows \u00d7 21 columns
Next we can create a StreamPipesFunction. For this we need to implement the 3 following methods:
onServiceStarted is called when the function gets started. There you can use the given meta information of the FunctionContext to initialize the function.onEvent is called when ever a new event arrives. The event contains the live data and you can use the streamId to identify a stream if the function is connected to multiple data streams.onServiceStopped is called when the function gets stopped.For this tutorial we just create a function that saves every new event in a pandas DataFrame and plots the first column of the DataFrame when the function gets stopped.
(If you want to use the same structure as in Java you can overwrite the getFunctionId and requiredStreamIds methods instead of using the FunctionDefinition)
from typing import Dict, Any\nimport pandas as pd\nfrom datetime import datetime\nimport matplotlib.pyplot as plt\nfrom streampipes.functions.function_handler import FunctionHandler\nfrom streampipes.functions.registration import Registration\nfrom streampipes.functions.streampipes_function import StreamPipesFunction\nfrom streampipes.functions.utils.function_context import FunctionContext\nfrom streampipes.model.resource.function_definition import FunctionDefinition, FunctionId\n\nclass ExampleFunction(StreamPipesFunction):\n def __init__(self, function_definition: FunctionDefinition) -> None:\n super().__init__(function_definition)\n # Create the Dataframe to save the live data\n self.df = pd.DataFrame()\n\n def onServiceStarted(self, context: FunctionContext):\n # Get the name of the timestamp field\n for event_property in context.schema[context.streams[0]].event_schema.event_properties:\n if event_property.property_scope == \"HEADER_PROPERTY\":\n self.timestamp = event_property.runtime_name\n\n def onEvent(self, event: Dict[str, Any], streamId: str):\n # Convert the unix timestamp to datetime\n event[self.timestamp] = datetime.fromtimestamp(event[self.timestamp] / 1000)\n # Add every value of the event to the DataFrame\n self.df = pd.concat(\n [self.df, pd.DataFrame({key: [event[key]] for key in event.keys()}).set_index(self.timestamp)]\n )\n\n def onServiceStopped(self):\n # Plot the first column of the Dataframe\n plt.figure(figsize=(10, 5))\n plt.xlabel(self.timestamp)\n plt.ylabel(self.df.columns[0])\n plt.plot(self.df.iloc[:, 0])\n plt.show()\n from typing import Dict, Any import pandas as pd from datetime import datetime import matplotlib.pyplot as plt from streampipes.functions.function_handler import FunctionHandler from streampipes.functions.registration import Registration from streampipes.functions.streampipes_function import StreamPipesFunction from streampipes.functions.utils.function_context import FunctionContext from streampipes.model.resource.function_definition import FunctionDefinition, FunctionId class ExampleFunction(StreamPipesFunction): def __init__(self, function_definition: FunctionDefinition) -> None: super().__init__(function_definition) # Create the Dataframe to save the live data self.df = pd.DataFrame() def onServiceStarted(self, context: FunctionContext): # Get the name of the timestamp field for event_property in context.schema[context.streams[0]].event_schema.event_properties: if event_property.property_scope == \"HEADER_PROPERTY\": self.timestamp = event_property.runtime_name def onEvent(self, event: Dict[str, Any], streamId: str): # Convert the unix timestamp to datetime event[self.timestamp] = datetime.fromtimestamp(event[self.timestamp] / 1000) # Add every value of the event to the DataFrame self.df = pd.concat( [self.df, pd.DataFrame({key: [event[key]] for key in event.keys()}).set_index(self.timestamp)] ) def onServiceStopped(self): # Plot the first column of the Dataframe plt.figure(figsize=(10, 5)) plt.xlabel(self.timestamp) plt.ylabel(self.df.columns[0]) plt.plot(self.df.iloc[:, 0]) plt.show() Now we can start the function. First we create an instance of the ExampleFunction and insert the element_id of the stream which data we want to consume. Then we have to register this function and we can start all functions by initializing the FunctionHandler. (it's also possible to register multiple functions with .register(...).register(...))
example_function = ExampleFunction(\n FunctionDefinition(\n function_id=FunctionId(id=\"example-function\"),\n consumed_streams=[\"urn:streampipes.apache.org:eventstream:uPDKLI\"]\n )\n)\n\nregistration = Registration()\nregistration.register(example_function)\n\nfunction_handler = FunctionHandler(registration, client)\nfunction_handler.initializeFunctions()\nexample_function = ExampleFunction( FunctionDefinition( function_id=FunctionId(id=\"example-function\"), consumed_streams=[\"urn:streampipes.apache.org:eventstream:uPDKLI\"] ) ) registration = Registration() registration.register(example_function) function_handler = FunctionHandler(registration, client) function_handler.initializeFunctions()
2022-12-14 10:43:42,810 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:153] [_make_request] - Successfully retrieved all resources.\n2022-12-14 10:43:42,812 - streampipes.functions.function_handler - [INFO] - [function_handler.py:82] [initializeFunctions] - Using NatsBroker for <__main__.ExampleFunction object at 0x000001F2EF298D90>\n
The DataFrame of the ExampleFunction gets updated whenever a new event arrives. Let's test this by executing the cell below.
import asyncio\nfrom asyncio.exceptions import CancelledError\nfrom IPython.display import clear_output\n\nwhile True:\n clear_output(wait=True)\n display(example_function.df)\n try:\n await asyncio.sleep(1)\n except CancelledError:\n break\nimport asyncio from asyncio.exceptions import CancelledError from IPython.display import clear_output while True: clear_output(wait=True) display(example_function.df) try: await asyncio.sleep(1) except CancelledError: break mass_flow density volume_flow sensor_fault_flags temperature sensorId timestamp 2022-12-14 10:43:43.357 10.955496 47.546290 1.001985 False 44.993413 flowrate02 2022-12-14 10:43:44.371 6.499040 44.392069 2.034402 False 41.232352 flowrate02 2022-12-14 10:43:45.382 10.168300 41.192146 9.724287 False 46.812779 flowrate02 2022-12-14 10:43:46.395 10.849059 50.086308 5.832691 False 45.860412 flowrate02 2022-12-14 10:43:47.410 3.081855 47.254246 8.860531 False 50.505801 flowrate02 ... ... ... ... ... ... ... 2022-12-14 10:44:43.920 1.803572 41.978894 10.294002 False 47.820239 flowrate02 2022-12-14 10:44:44.932 1.967062 42.212883 3.237440 False 49.047258 flowrate02 2022-12-14 10:44:45.934 4.457819 47.561256 0.315024 False 40.223413 flowrate02 2022-12-14 10:44:46.949 8.745343 46.346891 7.439090 False 41.982529 flowrate02 2022-12-14 10:44:47.950 5.828744 47.679720 6.307405 False 42.100354 flowrate02
65 rows \u00d7 6 columns
The while loop just displays the the DataFrame every second until the cell is stopped. We could achieve the same result manually by executing example_function.df repeatedly.
You can stop the functions whenever you want by executing the command below.
In\u00a0[\u00a0]: Copied!function_handler.disconnect()\nfunction_handler.disconnect()
2022-12-14 10:44:53,309 - streampipes.functions.broker.nats_broker - [INFO] - [nats_broker.py:67] [disconnect] - Stopped connection to stream: urn:streampipes.apache.org:eventstream:uPDKLI\n
That's enough for this tutorial. Now you can try to write your own StreamPipesFunction. All you need to do is creating a new class, implementing the 4 required methods and registering the function.
Want to see more exciting use cases you can achieve with StreamPipes functions in Python? Then don't hesitate and jump to our next tutorial on applying online machine learning algorithms to StreamPipes data streams with River.
How do you like this tutorial? We hope you like it and would love to receive some feedback from you. Just go to our GitHub discussion page and let us know your impression. We'll read and react to them all, we promise!
"},{"location":"tutorials/3-getting-live-data-from-the-streampipes-data-stream/#Getting-live-data-from-the-StreamPipes-data-stream","title":"Getting live data from the StreamPipes data stream\u00b6","text":"In the last tutorial (Extracting Data from the StreamPipes data lake) we learned how to extract the stored data from a StreamPipes data lake. This tutorial is about the StreamPipes data stream and shows how to get the live data from StreamPipes into Python. Therefore, we first create the client instance as before.
from streampipes.client import StreamPipesClient\nfrom streampipes.client.config import StreamPipesClientConfig\nfrom streampipes.client.credential_provider import StreamPipesApiKeyCredentials\nfrom streampipes.client import StreamPipesClient from streampipes.client.config import StreamPipesClientConfig from streampipes.client.credential_provider import StreamPipesApiKeyCredentials In\u00a0[\u00a0]: Copied!
# you can install all required dependencies for this tutorial by executing the following command\n%pip install river streampipes\n# you can install all required dependencies for this tutorial by executing the following command %pip install river streampipes In\u00a0[2]: Copied!
import os\n\nos.environ[\"SP_USERNAME\"] = \"admin@streampipes.apache.org\"\nos.environ[\"SP_API_KEY\"] = \"XXX\"\n\n# Use this if you work locally:\nos.environ[\"BROKER-HOST\"] = \"localhost\" \nos.environ[\"KAFKA-PORT\"] = \"9094\" # When using Kafka as message broker\nimport os os.environ[\"SP_USERNAME\"] = \"admin@streampipes.apache.org\" os.environ[\"SP_API_KEY\"] = \"XXX\" # Use this if you work locally: os.environ[\"BROKER-HOST\"] = \"localhost\" os.environ[\"KAFKA-PORT\"] = \"9094\" # When using Kafka as message broker In\u00a0[3]: Copied!
client_config = StreamPipesClientConfig(\n credential_provider=StreamPipesApiKeyCredentials(),\n host_address=\"localhost\",\n port=80,\n https_disabled=True,\n)\nclient = StreamPipesClient(client_config=client_config)\nclient_config = StreamPipesClientConfig( credential_provider=StreamPipesApiKeyCredentials(), host_address=\"localhost\", port=80, https_disabled=True, ) client = StreamPipesClient(client_config=client_config)
2023-01-27 16:04:24,784 - streampipes.client.client - [INFO] - [client.py:128] [_set_up_logging] - Logging successfully initialized with logging level INFO.\nIn\u00a0[4]: Copied!
client.dataStreamApi.all().to_pandas()\nclient.dataStreamApi.all().to_pandas()
2023-01-27 16:04:28,212 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:163] [_make_request] - Successfully retrieved all resources.\nOut[4]: element_id name description icon_url app_id includes_assets includes_locales internally_managed measurement_object index ... dom rev num_transport_protocols num_measurement_capability num_application_links num_included_assets num_connected_to num_category num_event_properties num_included_locales 0 sp:spdatastream:xboBFK Test None None False False True None 0 ... None 5-558c861debc745e1ebae29a266a8bdb9 1 0 0 0 0 0 7 0 1 urn:streampipes.apache.org:eventstream:Wgyrse Test File None None False False True None 0 ... None 4-66548b6b84287011b7cec0876ef82baf 1 0 0 0 0 0 2 0
2 rows \u00d7 22 columns
In\u00a0[5]: Copied!from river import cluster, compose, preprocessing\nfrom streampipes.function_zoo.river_function import OnlineML\nfrom streampipes.functions.utils.data_stream_generator import RuntimeType\n\nk_means = compose.Pipeline(\n (\"drop_features\", compose.Discard(\"sensorId\", \"timestamp\")),\n (\"scale\", preprocessing.StandardScaler()),\n (\"k_means\", cluster.KMeans(n_clusters=2)),\n)\n\nclustering = OnlineML(\n client=client, stream_ids=[\"sp:spdatastream:xboBFK\"], model=k_means, prediction_type=RuntimeType.INTEGER.value\n)\nclustering.start()\nfrom river import cluster, compose, preprocessing from streampipes.function_zoo.river_function import OnlineML from streampipes.functions.utils.data_stream_generator import RuntimeType k_means = compose.Pipeline( (\"drop_features\", compose.Discard(\"sensorId\", \"timestamp\")), (\"scale\", preprocessing.StandardScaler()), (\"k_means\", cluster.KMeans(n_clusters=2)), ) clustering = OnlineML( client=client, stream_ids=[\"sp:spdatastream:xboBFK\"], model=k_means, prediction_type=RuntimeType.INTEGER.value ) clustering.start()
2023-01-27 16:04:35,599 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:163] [_make_request] - Successfully retrieved all resources.\n2023-01-27 16:04:35,599 - streampipes.functions.function_handler - [INFO] - [function_handler.py:64] [initializeFunctions] - Create output data stream \"sp:spdatastream:cwKPoo\" for the function \"65cf8b86-bcdf-433e-a1c7-3e920eab55d0\"\n2023-01-27 16:04:37,766 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:163] [_make_request] - Successfully retrieved all resources.\n2023-01-27 16:04:37,767 - streampipes.functions.function_handler - [INFO] - [function_handler.py:78] [initializeFunctions] - Using NatsBroker for RiverFunction\n
2023-01-27 16:04:37,791 - streampipes.functions.broker.nats_broker - [INFO] - [nats_broker.py:48] [_makeConnection] - Connected to NATS at localhost:4222\n2023-01-27 16:04:37,791 - streampipes.functions.broker.nats_broker - [INFO] - [nats_broker.py:48] [_makeConnection] - Connected to NATS at localhost:4222\n2023-01-27 16:04:37,792 - streampipes.functions.broker.nats_broker - [INFO] - [nats_broker.py:58] [createSubscription] - Subscribed to stream: sp:spdatastream:xboBFK\nIn\u00a0[6]: Copied!
clustering.set_learning(False)\nclustering.set_learning(False) In\u00a0[\u00a0]: Copied!
clustering.stop()\nclustering.stop()
2023-01-27 16:04:57,303 - streampipes.functions.broker.nats_broker - [INFO] - [nats_broker.py:82] [disconnect] - Stopped connection to stream: sp:spdatastream:xboBFK\n2023-01-27 16:04:57,304 - streampipes.functions.broker.nats_broker - [INFO] - [nats_broker.py:82] [disconnect] - Stopped connection to stream: sp:spdatastream:cwKPoo\nIn\u00a0[\u00a0]: Copied!
import pickle\nfrom river import compose, tree\nfrom streampipes.function_zoo.river_function import OnlineML\nfrom streampipes.functions.utils.data_stream_generator import RuntimeType\n\nhoeffding_tree = compose.Pipeline(\n (\"drop_features\", compose.Discard(\"sensorId\", \"timestamp\")),\n (\"hoeffding_tree\", tree.HoeffdingTreeRegressor(grace_period=5)),\n)\n\n\ndef draw_tree(self, event, streamId):\n\"\"\"Draw the tree and save the image.\"\"\"\n if self.learning:\n if self.model[1].n_nodes != None:\n self.model[1].draw().render(\"hoeffding_tree\", format=\"png\", cleanup=True)\n\n\ndef save_model(self):\n\"\"\"Save the trained model.\"\"\"\n with open(\"hoeffding_tree.pkl\", \"wb\") as f:\n pickle.dump(self.model, f)\n\n\nregressor = OnlineML(\n client=client,\n stream_ids=[\"sp:spdatastream:xboBFK\"],\n model=hoeffding_tree,\n prediction_type=RuntimeType.FLOAT.value,\n supervised=True,\n target_label=\"temperature\",\n on_event=draw_tree,\n on_stop=save_model,\n)\nregressor.start()\nimport pickle from river import compose, tree from streampipes.function_zoo.river_function import OnlineML from streampipes.functions.utils.data_stream_generator import RuntimeType hoeffding_tree = compose.Pipeline( (\"drop_features\", compose.Discard(\"sensorId\", \"timestamp\")), (\"hoeffding_tree\", tree.HoeffdingTreeRegressor(grace_period=5)), ) def draw_tree(self, event, streamId): \"\"\"Draw the tree and save the image.\"\"\" if self.learning: if self.model[1].n_nodes != None: self.model[1].draw().render(\"hoeffding_tree\", format=\"png\", cleanup=True) def save_model(self): \"\"\"Save the trained model.\"\"\" with open(\"hoeffding_tree.pkl\", \"wb\") as f: pickle.dump(self.model, f) regressor = OnlineML( client=client, stream_ids=[\"sp:spdatastream:xboBFK\"], model=hoeffding_tree, prediction_type=RuntimeType.FLOAT.value, supervised=True, target_label=\"temperature\", on_event=draw_tree, on_stop=save_model, ) regressor.start() In\u00a0[9]: Copied!
regressor.set_learning(False)\nregressor.set_learning(False) In\u00a0[\u00a0]: Copied!
regressor.stop()\nregressor.stop() In\u00a0[\u00a0]: Copied!
import pickle\nfrom river import compose, tree\nfrom streampipes.function_zoo.river_function import OnlineML\nfrom streampipes.functions.utils.data_stream_generator import RuntimeType\n\ndecision_tree = compose.Pipeline(\n (\"drop_features\", compose.Discard(\"sensorId\", \"timestamp\")),\n (\"decision_tree\", tree.ExtremelyFastDecisionTreeClassifier(grace_period=5)),\n)\n\n\ndef draw_tree(self, event, streamId):\n\"\"\"Draw the tree and save the image.\"\"\"\n if self.learning:\n if self.model[1].n_nodes != None:\n self.model[1].draw().render(\"decicion_tree\", format=\"png\", cleanup=True)\n\n\ndef save_model(self):\n\"\"\"Save the trained model.\"\"\"\n with open(\"decision_tree.pkl\", \"wb\") as f:\n pickle.dump(self.model, f)\n\n\nclassifier = OnlineML(\n client=client,\n stream_ids=[\"sp:spdatastream:xboBFK\"],\n model=decision_tree,\n prediction_type=RuntimeType.BOOLEAN.value,\n supervised=True,\n target_label=\"sensor_fault_flags\",\n on_event=draw_tree,\n on_stop=save_model,\n)\nclassifier.start()\nimport pickle from river import compose, tree from streampipes.function_zoo.river_function import OnlineML from streampipes.functions.utils.data_stream_generator import RuntimeType decision_tree = compose.Pipeline( (\"drop_features\", compose.Discard(\"sensorId\", \"timestamp\")), (\"decision_tree\", tree.ExtremelyFastDecisionTreeClassifier(grace_period=5)), ) def draw_tree(self, event, streamId): \"\"\"Draw the tree and save the image.\"\"\" if self.learning: if self.model[1].n_nodes != None: self.model[1].draw().render(\"decicion_tree\", format=\"png\", cleanup=True) def save_model(self): \"\"\"Save the trained model.\"\"\" with open(\"decision_tree.pkl\", \"wb\") as f: pickle.dump(self.model, f) classifier = OnlineML( client=client, stream_ids=[\"sp:spdatastream:xboBFK\"], model=decision_tree, prediction_type=RuntimeType.BOOLEAN.value, supervised=True, target_label=\"sensor_fault_flags\", on_event=draw_tree, on_stop=save_model, ) classifier.start() In\u00a0[12]: Copied!
classifier.set_learning(False)\nclassifier.set_learning(False) In\u00a0[\u00a0]: Copied!
classifier.stop()\nclassifier.stop()
That's already it! Isn't it truly easy to apply Online ML with StreamPipes and River? Please go ahead and apply it to your own use cases. We would be happy to hear about them!
Want to see more exciting use cases you can achieve with StreamPipes functions in Python? Then don\u2019t hesitate and jump to our next tutorial on using interoperable machine learning algorithm models with StreamPipes Python and ONNX.
How do you like this tutorial? We hope you like it and would love to receive some feedback from you. Just go to our GitHub discussion page and let us know your impression. We'll read and react to them all, we promise!
"},{"location":"tutorials/4-using-online-machine-learning-on-a-streampipes-data-stream/#Using-Online-Machine-Learning-on-a-StreamPipes-data-stream","title":"Using Online Machine Learning on a StreamPipes data stream\u00b6","text":"The last tutorial (Getting live data from the StreamPipes data stream) showed how we can connect to a data stream, and it would be possible to use Online Machine Learning with this approach and train a model with the incoming events at the onEvent method. However, the StreamPipes client also provides an easier way to do this with the use of the River library for Online Machine Learning. We will have a look at this now.
After we configured the client as usual, we can start with the new part. The approach is straight forward and you can start with the ML part in just 3 steps:
OnlineML wrapper to fit to your model and insert the client and required data stream ids.A StreamPipesFunction is then started, which trains the model for each new event. It also creates an output data stream which will send the prediction of the model back to StreamPipes. This output stream can be seen when creating a new pipeline and can be used like every other data source. So you can use it in a pipeline and save the predictions in a Data Lake. You can also stop and start the training with the method set_learning. To stop the whole function use the stop methode and if you want to delete the output stream entirely, you can go to the Pipeline Element Installer in StreamPipes and uninstall it.
Now let's take a look at some examples. If you want to execute the examples below you have to create an adapter for the Machine Data Simulator, select the flowrate sensor and insert the stream id of this stream.
%pip install git+https://github.com/apache/streampipes.git#subdirectory=streampipes-client-python\n%pip install scikit-learn==1.4.0 skl2onnx==1.16.0 onnxruntime==1.17.1 matplotlib==3.8.3\n%pip install git+https://github.com/apache/streampipes.git#subdirectory=streampipes-client-python %pip install scikit-learn==1.4.0 skl2onnx==1.16.0 onnxruntime==1.17.1 matplotlib==3.8.3 In\u00a0[8]: Copied!
import os\nfrom streampipes.client import StreamPipesClient\nfrom streampipes.client.config import StreamPipesClientConfig\nfrom streampipes.client.credential_provider import StreamPipesApiKeyCredentials\n\nos.environ[\"BROKER-HOST\"] = \"localhost\"\nos.environ[\"KAFKA-PORT\"] = \"9094\" # When using Kafka as message broker\n\nconfig = StreamPipesClientConfig(\n credential_provider=StreamPipesApiKeyCredentials(\n username=\"admin@streampipes.apache.org\",\n api_key=\"TOKEN\",\n ),\n host_address=\"localhost\",\n https_disabled=True,\n port=80\n)\n\nclient = StreamPipesClient(client_config=config)\nimport os from streampipes.client import StreamPipesClient from streampipes.client.config import StreamPipesClientConfig from streampipes.client.credential_provider import StreamPipesApiKeyCredentials os.environ[\"BROKER-HOST\"] = \"localhost\" os.environ[\"KAFKA-PORT\"] = \"9094\" # When using Kafka as message broker config = StreamPipesClientConfig( credential_provider=StreamPipesApiKeyCredentials( username=\"admin@streampipes.apache.org\", api_key=\"TOKEN\", ), host_address=\"localhost\", https_disabled=True, port=80 ) client = StreamPipesClient(client_config=config)
2024-03-26 10:21:38,538 - streampipes.client.client - [INFO] - [client.py:198] [_set_up_logging] - Logging successfully initialized with logging level INFO.\n2024-03-26 10:21:38,632 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:164] [_make_request] - Successfully retrieved all resources.\n2024-03-26 10:21:38,634 - streampipes.client.client - [INFO] - [client.py:171] [_get_server_version] - The StreamPipes version was successfully retrieved from the backend: 0.95.0. By means of that, authentication via the provided credentials is also tested successfully.\n
The main objective of this tutorial is to demonstrate how to make predictions with an existing and pre-trained ML model using a StreamPipes function and ONNX. Therefore, you can skip the following sections on use case and model training if you already have an existing ONNX model and are only interested in applying it using StreamPipes.
In\u00a0[9]: Copied!flowrate_df = client.dataLakeMeasureApi.get(\"flow-rate\").to_pandas()\nX = flowrate_df[\"volume_flow\"].values.reshape(-1, 1).astype(\"float32\")\nflowrate_df = client.dataLakeMeasureApi.get(\"flow-rate\").to_pandas() X = flowrate_df[\"volume_flow\"].values.reshape(-1, 1).astype(\"float32\")
2024-03-26 10:21:48,582 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:164] [_make_request] - Successfully retrieved all resources.\n
Let's fit the model to the data:
In\u00a0[14]: Copied!from sklearn.ensemble import IsolationForest\n\nmodel = IsolationForest(contamination=0.01)\nmodel.fit(X)\nfrom sklearn.ensemble import IsolationForest model = IsolationForest(contamination=0.01) model.fit(X) Out[14]:
IsolationForest(contamination=0.01)In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.\u00a0\u00a0IsolationForest?Documentation for IsolationForestiFitted
IsolationForest(contamination=0.01)
The contamination parameter models the proportion of outliers in the data. See the scikit-learn documentation for more information.
Before we convert the model to an ONNX representation, let's do a quick visual analysis of the model results:
In\u00a0[15]: Copied!import matplotlib.pyplot as plt\n\nflowrate_df[\"anomaly\"] = model.predict(X)\n\nfig, ax = plt.subplots(figsize=(10, 6))\nanomalies = flowrate_df.loc[flowrate_df[\"anomaly\"] == -1, [\"volume_flow\"]]\nax.plot(flowrate_df.index, flowrate_df['volume_flow'], color='black', label='volume_flow')\nax.scatter(anomalies.index, anomalies['volume_flow'], color='red', label='Anomaly')\nplt.legend()\nplt.show()\nimport matplotlib.pyplot as plt flowrate_df[\"anomaly\"] = model.predict(X) fig, ax = plt.subplots(figsize=(10, 6)) anomalies = flowrate_df.loc[flowrate_df[\"anomaly\"] == -1, [\"volume_flow\"]] ax.plot(flowrate_df.index, flowrate_df['volume_flow'], color='black', label='volume_flow') ax.scatter(anomalies.index, anomalies['volume_flow'], color='red', label='Anomaly') plt.legend() plt.show()
Okay, that looks quite reasonable so lets covert the model to an ONNX representation so that we can make use of it later.
In\u00a0[16]: Copied!from onnxconverter_common import FloatTensorType\nfrom skl2onnx import to_onnx\n\nmodel_onnx = to_onnx(\n model,\n initial_types=[('input', FloatTensorType([None, X.shape[1]]))],\n target_opset={'ai.onnx.ml': 3, 'ai.onnx': 15, '': 15}\n)\n\nwith open(\"isolation_forest.onnx\", \"wb\") as f:\n f.write(model_onnx.SerializeToString())\n from onnxconverter_common import FloatTensorType from skl2onnx import to_onnx model_onnx = to_onnx( model, initial_types=[('input', FloatTensorType([None, X.shape[1]]))], target_opset={'ai.onnx.ml': 3, 'ai.onnx': 15, '': 15} ) with open(\"isolation_forest.onnx\", \"wb\") as f: f.write(model_onnx.SerializeToString()) In\u00a0[23]: Copied! import numpy as np\nimport onnxruntime as rt\n\nfrom streampipes.functions.broker.broker_handler import get_broker_description\nfrom streampipes.functions.streampipes_function import StreamPipesFunction\nfrom streampipes.functions.utils.data_stream_generator import create_data_stream, RuntimeType\nfrom streampipes.functions.utils.function_context import FunctionContext\nfrom streampipes.model.resource import FunctionDefinition, DataStream\n\nfrom typing import Dict, Any, List\n\n\nclass ONNXFunction(StreamPipesFunction):\n\n def __init__(self, feature_names: list[str], input_stream: DataStream):\n output_stream = create_data_stream(\n name=\"flowrate-prediction\",\n attributes={\n \"is_anomaly\": RuntimeType.BOOLEAN.value\n },\n broker=get_broker_description(input_stream)\n )\n\n function_definition = FunctionDefinition(\n consumed_streams=[input_stream.element_id]\n ).add_output_data_stream(output_stream)\n\n self.feature_names = feature_names\n self.input_name = None\n self.output_name = None\n self.session = None\n\n super().__init__(function_definition=function_definition)\n\n def onServiceStarted(self, context: FunctionContext) -> None:\n self.session = rt.InferenceSession(\n path_or_bytes=\"isolation_forest.onnx\",\n providers=rt.get_available_providers(),\n )\n self.input_name = self.session.get_inputs()[0].name\n self.output_name = self.session.get_outputs()[0].name\n\n def onEvent(self, event: Dict[str, Any], streamId: str) -> None:\n feature_vector = []\n for feature in self.feature_names:\n feature_vector.append(event[feature])\n\n prediction = self.session.run(\n [self.output_name],\n {self.input_name: np.expand_dims(np.array(feature_vector), axis=0).astype(\"float32\")}\n )[0]\n\n output = {\n \"is_anomaly\": int(prediction[0]) == -1\n }\n\n self.add_output(\n stream_id=self.function_definition.get_output_stream_ids()[0],\n event=output\n )\n\n def onServiceStopped(self) -> None:\n pass\n import numpy as np import onnxruntime as rt from streampipes.functions.broker.broker_handler import get_broker_description from streampipes.functions.streampipes_function import StreamPipesFunction from streampipes.functions.utils.data_stream_generator import create_data_stream, RuntimeType from streampipes.functions.utils.function_context import FunctionContext from streampipes.model.resource import FunctionDefinition, DataStream from typing import Dict, Any, List class ONNXFunction(StreamPipesFunction): def __init__(self, feature_names: list[str], input_stream: DataStream): output_stream = create_data_stream( name=\"flowrate-prediction\", attributes={ \"is_anomaly\": RuntimeType.BOOLEAN.value }, broker=get_broker_description(input_stream) ) function_definition = FunctionDefinition( consumed_streams=[input_stream.element_id] ).add_output_data_stream(output_stream) self.feature_names = feature_names self.input_name = None self.output_name = None self.session = None super().__init__(function_definition=function_definition) def onServiceStarted(self, context: FunctionContext) -> None: self.session = rt.InferenceSession( path_or_bytes=\"isolation_forest.onnx\", providers=rt.get_available_providers(), ) self.input_name = self.session.get_inputs()[0].name self.output_name = self.session.get_outputs()[0].name def onEvent(self, event: Dict[str, Any], streamId: str) -> None: feature_vector = [] for feature in self.feature_names: feature_vector.append(event[feature]) prediction = self.session.run( [self.output_name], {self.input_name: np.expand_dims(np.array(feature_vector), axis=0).astype(\"float32\")} )[0] output = { \"is_anomaly\": int(prediction[0]) == -1 } self.add_output( stream_id=self.function_definition.get_output_stream_ids()[0], event=output ) def onServiceStopped(self) -> None: pass Let's dive a little deeper into the different parts of the function
__init__: First, we need to take care of the data stream that is required to send the predictions from our function to StreamPipes. Thus, we create a dedicated output data stream which we need to provide with the attributes our event will consist of (a timestamp attribute is always added automatically). This output data stream needs to be registered at the function definition which is to be passed to the parent class. Lastly, we need to define some instance variables that are mainly required for the ONNX runtime.
onServiceStarted: Here we prepare the ONNX runtime session by creating an InferenceSession and retrieving the corresponding configuration parameters.
onEvent: Following the parameter names specified by self.feature_names, we extract all feature values from the current event. Subsequently, the corresponding feature vector is transmitted to the ONNX runtime session. The resulting prediction is then converted into our output event, where a value of -1 signifies an anomaly. Finally, the generated output event is forwarded to StreamPipes.
Having the function code in place, we can start the function with the following:
In\u00a0[25]: Copied!from streampipes.functions.registration import Registration\nfrom streampipes.functions.function_handler import FunctionHandler\n\nstream = [\n stream\n for stream\n in client.dataStreamApi.all()\n if stream.name == \"flow-rate\"\n][0]\n\nfunction = ONNXFunction(\n feature_names=[\"volume_flow\"],\n input_stream=stream\n)\n\nregistration = Registration()\nregistration.register(function)\nfunction_handler = FunctionHandler(registration, client)\nfunction_handler.initializeFunctions()\nfrom streampipes.functions.registration import Registration from streampipes.functions.function_handler import FunctionHandler stream = [ stream for stream in client.dataStreamApi.all() if stream.name == \"flow-rate\" ][0] function = ONNXFunction( feature_names=[\"volume_flow\"], input_stream=stream ) registration = Registration() registration.register(function) function_handler = FunctionHandler(registration, client) function_handler.initializeFunctions()
2024-03-26 12:39:50,443 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:164] [_make_request] - Successfully retrieved all resources.\n2024-03-26 12:39:50,502 - streampipes.functions.function_handler - [INFO] - [function_handler.py:76] [initializeFunctions] - The data stream could not be created.\n2024-03-26 12:39:50,503 - streampipes.functions.function_handler - [INFO] - [function_handler.py:78] [initializeFunctions] - This is due to the fact that this data stream already exists. Continuing with the existing data stream.\n2024-03-26 12:39:50,503 - streampipes.functions.function_handler - [INFO] - [function_handler.py:84] [initializeFunctions] - Using output data stream 'sp:spdatastream:flowrate-prediction' for function '7c06fa31-9534-4f91-9c50-b7a3607ec3dc'\n2024-03-26 12:39:50,548 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:164] [_make_request] - Successfully retrieved all resources.\n2024-03-26 12:39:50,549 - streampipes.functions.function_handler - [INFO] - [function_handler.py:100] [initializeFunctions] - Using KafkaConsumer for ONNXFunction\n
We can now access the live values of the prediction in the StreamPipes UI, e.g., in the pipeline editor.
That's already it. We hope this tutorial serves as an illustration how ML models can be utilized in StreamPipes with the help of ONNX.
How do you like this tutorial? We hope you like it and would love to receive some feedback from you. Just go to our GitHub discussion page and let us know your impression. We'll read and react to them all, we promise!
"},{"location":"tutorials/5-applying-interoperable-machine-learning-in-streampipes/#Applying-Interoperable-Machine-Learning-in-StreamPipes","title":"Applying Interoperable Machine Learning in StreamPipes\u00b6","text":"The last tutorial (Using Online Machine Learning on a StreamPipes data stream) demonstrated how patterns in streaming data can be learned online. In contrast, this tutorial demonstrates how one can apply a pre-trained machine learning (ML) model to a StreamPipes data stream making use of ONNX. We will show how StreamPipes can be used for both: extracting historical data for training purposes and using model inference on live data with a pre-trained model.
"},{"location":"tutorials/5-applying-interoperable-machine-learning-in-streampipes/#Preparation","title":"Preparation\u00b6","text":"The following lines configure the client and establish a connection to the StreamPipes instance. If you're not familiar with it or anything is unclear, please have a look at our first tutorial.
"},{"location":"tutorials/5-applying-interoperable-machine-learning-in-streampipes/#Machine-Learning-Use-Case","title":"Machine Learning Use Case\u00b6","text":"In this tutorial, we will use data generated by the Machine Data Simulator adapter. More specifically, we will focus on the flowrate data, which consists of various sensor values coming from a water pipe system. Our goal is keep an eye on the parameter volume_flow, which represents the current volume flow in cubic meters/second. For this parameter, we want to detect anomalies that could indicate problems such as leaks, blockages, etc.
To get the concerned data, we simply need to create an instance of the machine data simulator and persist the data in the data lake:
If you choose to perform the model training step yourself, you will need to wait approximately 15 minutes for enough data to be available for model training. If you want to speed this up, you can configure a lower wait time when creating the adapter. Please be aware that this also influences the inference scenario.
"},{"location":"tutorials/5-applying-interoperable-machine-learning-in-streampipes/#Model-Training-with-Historic-Data","title":"Model Training with Historic Data\u00b6","text":"As said above, the aim of our model is to detect anomalies of the volume_flow parameter. For this task, we will use Isolation Forests. Please note that the focus of the tutorial is not on training the model, so please be patient even though the training is very simplified and lacks important preparation steps such as standardization.
As a first step, lets query the flowrate data from the StreamPipes data lake and extract the values of volume_flow as a feature:
Utilizing a pre-trained model within StreamPipes becomes seamless with the ONNX interoperability standard, enabling effortless application of your existing model on live data streams.
Interacting with live data from StreamPipes is facilitated through StreamPipes functions. Below, we'll create a Python StreamPipes function that leverages an ONNX model to generate predictions for each incoming event, making the results accessible as a data stream within StreamPipes for subsequent steps.
So let's create an ONNXFunction that is capable of applying a model in ONNX representation to a StreamPipes data stream. If you'd like to read more details about how functions are defined, refer to our third tutorial.
Apache StreamPipes meets Python! We are working highly motivated on a Python library to interact with StreamPipes. In this way, we would like to unite the power of StreamPipes to easily connect to and read from different data sources, especially in the IoT domain, and the amazing universe of data analytics libraries in Python.
"},{"location":"#quickstart","title":"\u26a1\ufe0f Quickstart","text":"As a quick example, we demonstrate how to set up and configure a StreamPipes client. In addition, we will get the available data lake measures out of StreamPipes.
from streampipes.client import StreamPipesClient\nfrom streampipes.client.config import StreamPipesClientConfig\nfrom streampipes.client.credential_provider import StreamPipesApiKeyCredentials\n\nconfig = StreamPipesClientConfig(\n credential_provider = StreamPipesApiKeyCredentials(\n username = \"test@streampipes.apache.org\",\n api_key = \"DEMO-KEY\",\n ),\n host_address = \"localhost\",\n https_disabled = True,\n port = 80\n)\n\nclient = StreamPipesClient(client_config=config)\n\n# get all available datat lake measures\nmeasures = client.dataLakeMeasureApi.all()\n\n# get amount of retrieved measures\nlen(measures)\n1\n# inspect the data lake measures as pandas dataframe\nmeasures.to_pandas()\nOutput:
measure_name timestamp_field ... pipeline_is_running num_event_properties\n0 test s0::timestamp ... False 2\n[1 rows x 6 columns]\nfrom streampipes.client.credential_provider import StreamPipesApiKeyCredentials\n\nStreamPipesApiKeyCredentials()\nSP_API_KEY and SP_USERNAME username is always the username that is used to log in into StreamPipes.
The api_key can be generated within the UI as demonstrated below:
This document describes how to easily set up your local dev environment to work on StreamPipes Python \ud83d\udc0d.
"},{"location":"getting-started/developing/#first-steps","title":"\ud83d\ude80 First Steps","text":"1) Set up your Python environment
Create a virtual Python environment using a tool of your choice. To manage dependencies, we use Poetry, so please install poetry in your local environment, e.g. via
pip install poetry\nOnce poetry is installed you can simply finalize your Python environment by running:
poetry install --with dev,stubs # install everything that is required for the development\npoetry install --with docs # install everything to work with the documentation\npoetry install --with dev,stubs,docs # install all optional dependencies related to development\n2) Install pre-commit hook
The pre-commit hook is run before every commit and takes care about code style, linting, type hints, import sorting, etc. It will stop your commit in case the changes do not apply the expected format. Always check to have the recent version of the pre-commit hook installed otherwise the CI build might fail. If you are interested, you can have a deeper look on the underlying library: pre-commit.
pre-commit install\nBelow we list some conventions that we have agreed on for creating StreamPipes Python. Please comply to them when you plan to contribute to this project. If you have any other suggestions or would like to discuss them, we would be happy to hear from you on our mailing list dev@streampipes.apache.org or in our discussions on GitHub.
1) Use numpy style for Python docstrings \ud83d\udcc4 Please stick to the numpy style when writing docstrings, as we require this for generating our documentation.
2) Provide tests \u2705 We are aiming for broad test coverage for the Python package and have therefore set a requirement of at least 90% unit test coverage. Therefore, please remember to write (unit) tests already during development. If you have problems with writing tests, don't hesitate to ask us for help directly in the PR or even before that via our mailing list (see above).
3) Build a similar API as the Java client provides \ud83d\udd04 Whenever possible, please try to develop the API of the Python library the same as the Java client or Java SDK. By doing so, we would like to provide a consistent developer experience and the basis for automated testing in the future.
"},{"location":"getting-started/developing/#dependency-management","title":"\ud83d\udce6 Dependency Management","text":"In case you want to add a new dependency to StreamPipes you can use the following command:
poetry add <dep-name>\nIf the dependency is only required for development purpose or the documentation, please stick to one the following:
poetry add <dep-name> --group dev\npoetry add <dep-name> --group stubs\npoetry add <dep-name> --group docs\nIn case you want to regenerate the poetry lock file, e.g., in case you manually updated the pyproject.toml, the following command should be used:
poetry lock --no-update\nAfter that, you should install the current version of the poetry lock file to keep your local environment consistent (see command above.)
"},{"location":"getting-started/developing/#documentation","title":"\ud83d\udcdaDocumentation","text":"To build our documentation, we use Materials for MkDocs. All files can be found within the docs directory. To pre-view your local version of the documentation, you can use the following command:
make livedoc\nBroadly speaking, we plan to expand or add new aspects/functionality to the library where we are focusing on the following:
In case you want to have a more detailed look on what we are currently planning, have a look at our open issues(more short-term driven).
Of course, contributions are always highly appreciated \ud83d\udd2e
Stay tuned!
"},{"location":"getting-started/developing/#contributing","title":"\ud83d\udc68\u200d\ud83d\udcbb Contributing","text":"Before opening a pull request, review the Get Involved page. It lists information that is required for contributing to StreamPipes.
When you contribute code, you affirm that the contribution is your original work and that you license the work to the project under the project's open source license. Whether or not you state this explicitly, by submitting any copyrighted material via pull request, email, or other means you agree to license the material under the project's open source license and warrant that you have the legal authority to do so.
"},{"location":"getting-started/first-steps/","title":"First Steps","text":""},{"location":"getting-started/first-steps/#installation","title":"\ud83d\udcda Installation","text":"The StreamPipes Python library is meant to work with Python 3.8 and above. Installation can be done via pip: You can install the latest development version from GitHub, as so:
pip install streampipes\n\n# if you want to have the current development state you can also execute\npip install git+https://github.com/apache/streampipes.git#subdirectory=streampipes-client-python\n# the corresponding documentation can be found here: https://streampipes.apache.org/docs/docs/python/dev/\nWhen working with the StreamPipes Python library it is inevitable to have a running StreamPipes instance to connect and interact with. In case you don't have a running instance at hand, you can easily set up one on your local machine. Hereby you need to consider that StreamPipes supports different message broker (e.g., Kafka, NATS). We will demonstrate below how you can easily set up StreamPipes for both supported message brokers.
"},{"location":"getting-started/first-steps/#start-streampipes-via-docker-compose","title":"\ud83d\udc33 Start StreamPipes via Docker Compose","text":"The easiest and therefore recommend way to get StreamPipes started is by using docker compose. Therefore, you need Docker running. You can check if Docker is ready on your machine by executing.
docker ps\nCONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES\n... ... ... ... ... ... ...\ndocker compose here."},{"location":"getting-started/first-steps/#setup-streampipes-with-nats-as-message-broker","title":"Setup StreamPipes with NATS as message broker","text":"The following shows how you can set up a StreamPipes instance that uses NATS as messaging layer. So in this scenario, we will go with docker-compose.nats.yml. Thereby, when running locally, we need to add the following port mapping entry to services.nats.ports:
- 4222:4222\nAfter this modification is applied, StreamPipes can simply be started with this command:
docker-compose -f docker-compose.nats.yml up -d\nOnce all services are started, you can access StreamPipes via http://localhost.
Alternatively, you can use docker-compose.yml to start StreamPipes with Kafka as messaging layer. When running locally we have to modify services.kafka.environment and add the ports to services.kafka.ports:
environment:\nKAFKA_LISTENER_SECURITY_PROTOCOL_MAP: PLAINTEXT:PLAINTEXT,OUTSIDE:PLAINTEXT\nKAFKA_ADVERTISED_LISTENERS: PLAINTEXT://:9092,OUTSIDE://localhost:9094\nKAFKA_LISTENERS: PLAINTEXT://:9092,OUTSIDE://:9094\n...\nports:\n- 9094:9094\ndocker-compose -f docker-compose.yml up -d\nOnce all services are started, you can access StreamPipes via http://localhost.
In case you want to have more control over your StreamPipes setup, you might take a look at our deployment CLI.
Have fun discovering StreamPipes and our Python library \ud83d\ude80
"},{"location":"getting-started/quickstart/","title":"Quickstart","text":""},{"location":"getting-started/quickstart/#quickstart","title":"\u26a1\ufe0f Quickstart","text":"As a quick example, we demonstrate how to set up and configure a StreamPipes client. In addition, we will get the available data lake measures out of StreamPipes.
from streampipes.client import StreamPipesClient\nfrom streampipes.client.config import StreamPipesClientConfig\nfrom streampipes.client.credential_provider import StreamPipesApiKeyCredentials\n\nconfig = StreamPipesClientConfig(\n credential_provider = StreamPipesApiKeyCredentials(\n username = \"test@streampipes.apache.org\",\n api_key = \"DEMO-KEY\",\n ),\n host_address = \"localhost\",\n https_disabled = True,\n port = 80\n)\n\nclient = StreamPipesClient(client_config=config)\n\n# get all available datat lake measures\nmeasures = client.dataLakeMeasureApi.all()\n\n# get amount of retrieved measures\nlen(measures)\n1\n# inspect the data lake measures as pandas dataframe\nmeasures.to_pandas()\nOutput:
measure_name timestamp_field ... pipeline_is_running num_event_properties\n0 test s0::timestamp ... False 2\n[1 rows x 6 columns]\nfrom streampipes.client.credential_provider import StreamPipesApiKeyCredentials\n\nStreamPipesApiKeyCredentials()\nSP_API_KEY and SP_USERNAME username is always the username that is used to log in into StreamPipes. The api_key can be generated within the UI as demonstrated below:
Implementation of the StreamPipes client. The client is designed as the central point of interaction with the StreamPipes API and provides all functionalities to communicate with the API.
"},{"location":"reference/client/client/#streampipes.client.client.StreamPipesClient","title":"StreamPipesClient(client_config, logging_level=logging.INFO)","text":"The client to connect to StreamPipes.
This is the central point of contact with StreamPipes and provides all the functionalities to interact with it.
The client provides so-called \"endpoints\" each of which refers to an endpoint of the StreamPipes API, e.g. .dataLakeMeasureApi. An endpoint provides the actual methods to interact with StreamPipes API.
client_config Configures the client to connect properly to the StreamPipes instance.
TYPE: StreamPipesClientConfig
logging_level Influences the log messages emitted by the StreamPipesClient
TYPE: Optional[int] DEFAULT: INFO
dataLakeMeasureApi Instance of the data lake measure endpoint
TYPE: DataLakeMeasureEndpoint
dataStreamApi Instance of the data stream endpoint
TYPE: DataStreamEndpoint
AttributeError: In case an invalid configuration of the StreamPipesClientConfig is passed
Examples:
from streampipes.client import StreamPipesClient\nfrom streampipes.client.config import StreamPipesClientConfig\nfrom streampipes.client.credential_provider import StreamPipesApiKeyCredentials\nclient_config = StreamPipesClientConfig(\n credential_provider=StreamPipesApiKeyCredentials(\n username=\"test-user\",\n api_key=\"api-key\"\n ),\n host_address=\"localhost\",\n https_disabled=True\n)\nThe following way of instantiating a client instance is intended to be consistent with the StreamPipes Java client.
client = StreamPipesClient.create(client_config=client_config)\nIf you prefer a more pythonic way, you can simply write:
client = StreamPipesClient(client_config=client_config)\nTo interact with an endpoint:
data_lake_measures = client.dataLakeMeasureApi.all()\nTo inspect returned data as a pandas dataframe:
data_lake_measures.to_pandas()\n#\n# measure_name timestamp_field ... pipeline_is_running num_event_properties\n# 0 test s0::timestamp ... False 2\n# [1 rows x 6 columns]\nbase_api_path: str property","text":"Constructs the basic API URL from the given client_config.
base_api_path basic API path of the connected StreamPipes instance
TYPE: str
http_headers: Dict[str, str] property","text":"Returns the HTTP headers used for all requests.
The HTTP headers are composed of the authentication headers supplied by the credential provider and additional required headers (currently this is only the application header).
RETURNS DESCRIPTIONhttp_headers Header information for HTTP requests as string key-value pairs.
TYPE: Dict[str, str]
create(client_config, logging_level=logging.INFO) classmethod","text":"Returns an instance of the StreamPipesPythonClient.
Provides consistency to the StreamPipes Java client.
PARAMETER DESCRIPTIONclient_config Configures the client to connect properly to the StreamPipes instance.
TYPE: StreamPipesClientConfig
logging_level Influences the log messages emitted by the StreamPipesClient.
TYPE: int DEFAULT: INFO
StreamPipesClient"},{"location":"reference/client/client/#streampipes.client.client.StreamPipesClient.describe","title":"describe()","text":"Prints a short description of the connected StreamPipes instance and the available resources to the console.
RETURNS DESCRIPTION None Examples:
client.describe()\nHi there!\nYou are connected to a StreamPipes instance running at http://localhost:80.\nThe following StreamPipes resources are available with this client:\n6x DataStreams\n1x DataLakeMeasures\nConfiguration class for the StreamPipes client.
"},{"location":"reference/client/config/#streampipes.client.config.StreamPipesClientConfig","title":"StreamPipesClientConfig(credential_provider, host_address, https_disabled=False, port=80, additional_headers=dict()) dataclass","text":"Configure the StreamPipes client in accordance to the actual StreamPipes instance to connect to.
An instance is provided to the StreamPipesClient to configure it properly.
credential_provider Provides the credentials to authenticate with the StreamPipes API.
TYPE: CredentialProvider
host_address Host address of the StreamPipes instance to connect to. Should be provided without the protocol/scheme, e.g. as localhost or streampipes.xyz.
TYPE: str
https_disabled Determines whether https is used to connect to StreamPipes.
TYPE: Optional[bool] DEFAULT: False
port Specifies the port under which the StreamPipes API is available, e.g., 80 (with http) or 443 (with https)
TYPE: Optional[int] DEFAULT: 80
additional_headers Specifies additional HTTP headers that should be sent with each request, e.g., proxy headers
TYPE: Optional[Dict[str, str]] DEFAULT: dict()
Examples:
see StreamPipesClient
"},{"location":"reference/client/credential_provider/","title":"Credential provider","text":"Implementation of credential providers. A credential provider supplies the specified sort of credentials in the appropriate HTTP header format. The headers are then used by the client to connect to StreamPipes.
"},{"location":"reference/client/credential_provider/#streampipes.client.credential_provider.CredentialProvider","title":"CredentialProvider","text":" Bases: ABC
Abstract implementation of a credential provider. Must be inherited by all credential providers.
"},{"location":"reference/client/credential_provider/#streampipes.client.credential_provider.CredentialProvider.make_headers","title":"make_headers(http_headers=None)","text":"Creates the HTTP headers for the specific credential provider.
Concrete authentication headers must be defined in the implementation of a credential provider.
PARAMETER DESCRIPTIONhttp_headers Additional HTTP headers the generated headers are extended by.
TYPE: Optional[Dict[str, str]] DEFAULT: None
https_headers Dictionary with header information as string key-value pairs. Contains all pairs given as parameter plus the header pairs for authentication determined by the credential provider.
TYPE: Dict[str, str]
StreamPipesApiKeyCredentials(username=None, api_key=None)","text":" Bases: CredentialProvider
A credential provider that allows authentication via a StreamPipes API Token.
The required token can be generated via the StreamPipes UI (see the description on our start-page.
Both parameters can either be passed as arguments or remain unset. If they are not passed, they are retrieved from environment variables:
SP_USERNAME is expected to contain the usernameSP_API_KEY is expected to contain the API keyusername The username to which the API token is granted, e.g., demo-user@streampipes.apche.org. If not passed, the username is retrieved from environment variable SP_USERNAME.
TYPE: Optional[str] DEFAULT: None
api_key The StreamPipes API key as it is displayed in the UI. If not passed, the api key is retrieved from environment variable SP_API_KEY
TYPE: Optional[str] DEFAULT: None
Examples:
see StreamPipesClient
"},{"location":"reference/client/credential_provider/#streampipes.client.credential_provider.StreamPipesApiKeyCredentials.from_env","title":"from_env(username_env, api_key_env) classmethod","text":"DEPRECATED - use the class constructor instead
Returns an API key provider parameterized via environment variables.
PARAMETER DESCRIPTIONusername_env Name of the environment variable that contains the username
TYPE: str
api_key_env Name of the environment variable that contains the API key
TYPE: str
StreamPipesApiKeyCredentials RAISES DESCRIPTION KeyError If one of the environment variables is not defined
"},{"location":"reference/client/credential_provider/#streampipes.client.credential_provider.StreamPipesApiKeyCredentials.make_headers","title":"make_headers(http_headers=None)","text":"Creates the HTTP headers for the specific credential provider.
Concrete authentication headers must be defined in the implementation of a credential provider.
PARAMETER DESCRIPTIONhttp_headers Additional HTTP headers the generated headers are extended by.
TYPE: Optional[Dict[str, str]] DEFAULT: None
https_headers Dictionary with header information as string key-value pairs. Contains all pairs given as parameter plus the header pairs for authentication determined by the credential provider.
TYPE: Dict[str, str]
General implementation for an endpoint. Provided classes and assets are aimed to be used for developing endpoints. An endpoint provides all options to communicate with a dedicated part of StreamPipes in a handy way.
"},{"location":"reference/endpoint/endpoint/#streampipes.endpoint.endpoint.APIEndpoint","title":"APIEndpoint(parent_client)","text":" Bases: Endpoint
Abstract implementation of an API endpoint.
Serves as template for all endpoints of the StreamPipes API. By design, endpoints are only instantiated within the __init__ method of the StreamPipesClient.
all()","text":"Get all resources of this endpoint provided by the StreamPipes API.
Results are provided as an instance of a ResourceContainer that allows to handle the returned resources in a comfortable and pythonic way.
container Container element that bundles the returned resources
TYPE: ResourceContainer
build_url()","text":"Builds the endpoint's URL of the API path.
RETURNS DESCRIPTIONurl The URL of the endpoint
TYPE: str
get(identifier, **kwargs)","text":"Queries the specified resource from the API endpoint.
PARAMETER DESCRIPTIONidentifier The identifier of the resource to be queried.
TYPE: str
resource The specified resource as an instance of the corresponding model class.
TYPE: Resource
post(resource)","text":"Allows to post a resource to the StreamPipes API.
PARAMETER DESCRIPTIONresource The resource to be posted.
TYPE: Resource
None"},{"location":"reference/endpoint/endpoint/#streampipes.endpoint.endpoint.Endpoint","title":"Endpoint(parent_client)","text":" Bases: ABC
Abstract implementation of a StreamPipes endpoint.
Serves as template for all endpoints used for interaction with a StreamPipes instance. By design, endpoints are only instantiated within the __init__ method of the StreamPipesClient.
parent_client This parameter expects the instance of StreamPipesClient the endpoint is attached to.
TYPE: StreamPipesClient
MessagingEndpoint(parent_client)","text":" Bases: Endpoint
Abstract implementation of a StreamPipes messaging endpoint.
Serves as template for all endpoints used for interacting with the StreamPipes messaging layer directly. Therefore, they need to provide the functionality to talk with the broker system running in StreamPipes. By design, endpoints are only instantiated within the __init__ method of the StreamPipesClient.
broker: Broker property writable","text":"Defines the broker instance that is used to connect to StreamPipes' messaging layer.
This instance enables the client to authenticate to the broker used in the target StreamPipes instance, to consume messages from and to write messages to the broker.
RAISES DESCRIPTIONMessagingEndpointNotConfiguredError If the endpoint is used before the broker instance is set via configure()
broker The broker instance to be used to communicate with StreamPipes' messaging layer.
TYPE: Broker
configure(broker)","text":"Configures the message endpoint by setting the broker instance to be used.
This configuration step is required before the endpoint can be actually used. The based broker instance is passed to an internal property.
broker Broker instance that should be used for this endpoint
TYPE: Broker
None"},{"location":"reference/endpoint/exceptions/","title":"Exceptions","text":"Custom exceptions dedicated to the endpoints module.
"},{"location":"reference/endpoint/exceptions/#streampipes.endpoint.exceptions.MessagingEndpointNotConfiguredError","title":"MessagingEndpointNotConfiguredError(endpoint_name)","text":" Bases: Exception
Exception that indicates that an instance of a messaging endpoint has not been configured.
This error occurs when an instance of a messaging endpoint is used before the broker instance to be used is configured by passing it to the configure() method.
endpoint_name The name of the endpoint that caused the error
TYPE: str
Specific implementation of the StreamPipes API's data lake measure endpoints. This endpoint allows to consume data stored in StreamPipes' data lake.
"},{"location":"reference/endpoint/api/data_lake_measure/#streampipes.endpoint.api.data_lake_measure.DataLakeMeasureEndpoint","title":"DataLakeMeasureEndpoint(parent_client)","text":" Bases: APIEndpoint
Implementation of the DataLakeMeasure endpoint.
This endpoint provides an interface to all data stored in the StreamPipes data lake.
Consequently, it allows querying metadata about available data sets (see all() method). The metadata is returned as an instance of DataLakeMeasures.
In addition, the endpoint provides direct access to the data stored in the data laka by querying a specific data lake measure using the get() method.
Examples:
from streampipes.client import StreamPipesClient\nfrom streampipes.client.config import StreamPipesClientConfig\nfrom streampipes.client.credential_provider import StreamPipesApiKeyCredentials\nclient_config = StreamPipesClientConfig(\n credential_provider=StreamPipesApiKeyCredentials(username=\"test-user\", api_key=\"api-key\"),\n host_address=\"localhost\",\n port=8082,\n https_disabled=True\n)\nclient = StreamPipesClient.create(client_config=client_config)\n# get all existing data lake measures from StreamPipes\ndata_lake_measures = client.dataLakeMeasureApi.all()\n\n# let's take a look how many we got\nlen(data_lake_measures)\n5\n# Retrieve a specific data lake measure as a pandas DataFrame\nflow_rate_pd = client.dataLakeMeasureApi.get(identifier=\"flow-rate\").to_pandas()\nflow_rate_pd\n time density mass_flow sensorId sensor_fault_flags temperature volume_flow\n0 2023-02-24T16:19:41.472Z 50.872730 3.309556 flowrate02 False 44.448483 5.793138\n1 2023-02-24T16:19:41.482Z 47.186588 5.608580 flowrate02 False 40.322033 0.058015\n2 2023-02-24T16:19:41.493Z 46.735321 7.692881 flowrate02 False 49.239639 10.283526\n3 2023-02-24T16:19:41.503Z 40.169796 3.632898 flowrate02 False 49.933754 6.893441\n4 2023-02-24T16:19:41.513Z 49.635124 0.711260 flowrate02 False 50.106617 2.999871\n.. ... ... ... ... ... ... ...\n995 2023-02-24T16:19:52.927Z 50.057495 1.740114 flowrate02 False 46.558231 1.818237\n996 2023-02-24T16:19:52.94Z 41.038895 7.211723 flowrate02 False 48.048622 2.127493\n997 2023-02-24T16:19:52.952Z 45.837013 7.770180 flowrate02 False 48.188026 7.892062\n998 2023-02-24T16:19:52.965Z 43.389065 4.458602 flowrate02 False 48.280899 5.733892\n999 2023-02-24T16:19:52.977Z 44.056030 2.592060 flowrate02 False 47.505951 4.260697\nAs you can see, the returned amount of rows per default is 1000. We can modify this behavior by passing the limit paramter.
flow_rate_pd = client.dataLakeMeasureApi.get(identifier=\"flow-rate\", limit=10).to_pandas()\nlen(flow_rate_pd)\n10\nIf we are only interested in the values for density, columns allows us to select the columns to be returned:
flow_rate_pd = client.dataLakeMeasureApi.get(identifier=\"flow-rate\", columns='density', limit=3).to_pandas()\nflow_rate_pd\n time density\n0 2023-02-24T16:19:41.472Z 50.872730\n1 2023-02-24T16:19:41.482Z 47.186588\n2 2023-02-24T16:19:41.493Z 46.735321\nThis is only a subset of the available query parameters, find them at MeasurementGetQueryConfig.
"},{"location":"reference/endpoint/api/data_lake_measure/#streampipes.endpoint.api.data_lake_measure.DataLakeMeasureEndpoint.all","title":"all()","text":"Get all resources of this endpoint provided by the StreamPipes API.
Results are provided as an instance of a ResourceContainer that allows to handle the returned resources in a comfortable and pythonic way.
container Container element that bundles the returned resources
TYPE: ResourceContainer
build_url()","text":"Builds the endpoint's URL of the API path.
RETURNS DESCRIPTIONurl The URL of the endpoint
TYPE: str
get(identifier, **kwargs)","text":"Queries the specified data lake measure from the API.
By default, the maximum number of returned records is 1000. This behaviour can be influenced by passing the parameter limit with a different value (see MeasurementGetQueryConfig).
identifier The identifier of the data lake measure to be queried.
TYPE: str
**kwargs Keyword arguments can be used to provide additional query parameters. The available query parameters are defined by the MeasurementGetQueryConfig.
TYPE: Optional[Dict[str, Any]] DEFAULT: {}
measurement The specified data lake measure
TYPE: QueryResult
Examples:
see directly at DataLakeMeasureEndpoint.
"},{"location":"reference/endpoint/api/data_lake_measure/#streampipes.endpoint.api.data_lake_measure.DataLakeMeasureEndpoint.post","title":"post(resource)","text":"Allows to post a resource to the StreamPipes API.
PARAMETER DESCRIPTIONresource The resource to be posted.
TYPE: Resource
None"},{"location":"reference/endpoint/api/data_lake_measure/#streampipes.endpoint.api.data_lake_measure.MeasurementGetQueryConfig","title":"MeasurementGetQueryConfig","text":" Bases: BaseModel
Config class describing the parameters of the get() method for measurements.
This config class is used to validate the provided query parameters for the GET endpoint of measurements. Additionally, it takes care of the conversion to a proper HTTP query string. Thereby, parameter names are adapted to the naming of the StreamPipes API, for which Pydantic aliases are used.
ATTRIBUTE DESCRIPTIONcolumns A comma separated list of column names (e.g., time,value) If provided, the returned data only consists of the given columns. Please be aware that the column time as an index is always included.
TYPE: Optional[List[str]]
end_date Limits the queried data to only include data that is older than the specified time. In other words, any data that occurred after the end_date will not be included in the query results.
TYPE: Optional[datetime]
limit Amount of records returned at maximum (default: 1000) This needs to be at least 1
TYPE: Optional[int]
offset Offset to be applied to returned data This needs to be at least 0
TYPE: Optional[int]
order Ordering of query results Allowed values: ASC and DESC (default: ASC)
TYPE: Optional[str]
page_no Page number used for paging operation This needs to be at least 1
TYPE: Optional[int]
start_date Limits the queried data to only include data that is newer than the specified time. In other words, any data that occurred before the start_date will not be included in the query results.
TYPE: Optional[datetime]
Config","text":"Pydantic Config class
"},{"location":"reference/endpoint/api/data_lake_measure/#streampipes.endpoint.api.data_lake_measure.MeasurementGetQueryConfig.build_query_string","title":"build_query_string()","text":"Builds a HTTP query string for the config.
This method returns an HTTP query string for the invoking config. It follows the following structure ?param1=value1¶m2=value2.... This query string is not an entire URL, instead it needs to appended to an API path.
query_param_string HTTP query params string (?param1=value1¶m2=value2...)
TYPE: str
StreamPipesQueryValidationError","text":" Bases: Exception
A custom exception to be raised when the validation of query parameter causes an error.
"},{"location":"reference/endpoint/api/data_stream/","title":"Data stream","text":"Specific implementation of the StreamPipes API's data stream endpoints.
"},{"location":"reference/endpoint/api/data_stream/#streampipes.endpoint.api.data_stream.DataStreamEndpoint","title":"DataStreamEndpoint(parent_client)","text":" Bases: APIEndpoint
Implementation of the DataStream endpoint.
Consequently, it allows querying metadata about available data streams (see all() method). The metadata is returned as an instance of DataStreams.
Examples:
from streampipes.client import StreamPipesClient\nfrom streampipes.client.config import StreamPipesClientConfig\nfrom streampipes.client.credential_provider import StreamPipesApiKeyCredentials\n\nclient_config = StreamPipesClientConfig(\n credential_provider=StreamPipesApiKeyCredentials(username=\"test-user\", api_key=\"api-key\"),\n host_address=\"localhost\",\n port=8082,\n https_disabled=True\n)\nclient = StreamPipesClient.create(client_config=client_config)\n# let's get all existing data streams in StreamPipes\ndata_streams = client.dataStreamApi.all()\nlen(data_streams)\n2\nall()","text":"Get all resources of this endpoint provided by the StreamPipes API.
Results are provided as an instance of a ResourceContainer that allows to handle the returned resources in a comfortable and pythonic way.
container Container element that bundles the returned resources
TYPE: ResourceContainer
build_url()","text":"Builds the endpoint's URL of the API path.
RETURNS DESCRIPTIONurl The URL of the endpoint
TYPE: str
get(identifier, **kwargs)","text":"Queries the specified resource from the API endpoint.
PARAMETER DESCRIPTIONidentifier The identifier of the resource to be queried.
TYPE: str
resource The specified resource as an instance of the corresponding model class.
TYPE: Resource
post(resource)","text":"Allows to post a resource to the StreamPipes API.
PARAMETER DESCRIPTIONresource The resource to be posted.
TYPE: Resource
None"},{"location":"reference/endpoint/api/version/","title":"Version","text":"Specific implementation of the StreamPipes API's version endpoint.
"},{"location":"reference/endpoint/api/version/#streampipes.endpoint.api.version.VersionEndpoint","title":"VersionEndpoint(parent_client)","text":" Bases: APIEndpoint
Implementation of the Versions endpoint.
This endpoint provides metadata about the StreamPipes version of the connected instance. It only allows to apply the get() method with an empty string as identifier.
parent_client The instance of StreamPipesClient the endpoint is attached to.
TYPE: StreamPipesClient
Examples:
>>> from streampipes.client import StreamPipesClient\n>>> from streampipes.client.config import StreamPipesClientConfig\n>>> from streampipes.client.credential_provider import StreamPipesApiKeyCredentials\n>>> client_config = StreamPipesClientConfig(\n... credential_provider=StreamPipesApiKeyCredentials(username=\"test-user\", api_key=\"api-key\"),\n... host_address=\"localhost\",\n... port=8082,\n... https_disabled=True\n... )\n>>> client = StreamPipesClient.create(client_config=client_config)\n>>> client.versionApi.get(identifier=\"\").to_dict(use_source_names=False)\n{'backend_version': '0.92.0-SNAPSHOT'}\nall()","text":"Usually, this method returns information about all resources provided by this endpoint. However, this endpoint does not support this kind of operation.
RAISES DESCRIPTIONNotImplementedError This endpoint does not return multiple entries, therefore this method is not available.
"},{"location":"reference/endpoint/api/version/#streampipes.endpoint.api.version.VersionEndpoint.build_url","title":"build_url()","text":"Builds the endpoint's URL of the API path.
RETURNS DESCRIPTIONurl The URL of the endpoint
TYPE: str
get(identifier, **kwargs)","text":"Queries the resource from the API endpoint.
For this endpoint only one resource is available.
PARAMETER DESCRIPTIONidentifier Not supported by this endpoint, is set to an empty string.
TYPE: str
ValueError Non-empty identifier is not supported by this endpoint. Please set identifier to an empty string or None.
versions The specified resource as an instance of the corresponding model class(Version). # noqa: 501
TYPE: Version
post(resource)","text":"Usually, this method allows to create via this endpoint. Since the data represented by this endpoint is immutable, it does not support this kind of operation.
RAISES DESCRIPTIONNotImplementedError This endpoint does not allow for POST requests, therefore this method is not available.
"},{"location":"reference/function_zoo/river_function/","title":"River function","text":""},{"location":"reference/function_zoo/river_function/#streampipes.function_zoo.river_function.OnlineML","title":"OnlineML(client, stream_ids, model, prediction_type=RuntimeType.STRING.value, supervised=False, target_label=None, on_start=lambda self, context: None, on_event=lambda self, event, streamId: None, on_stop=lambda self: None)","text":"Wrapper class to enable an easy usage for Online Machine Learning models of the River library.
It creates a StreamPipesFunction to train a model with the incoming events of a data stream and creates an output data stream that publishes the prediction to StreamPipes.
PARAMETER DESCRIPTIONclient The client for the StreamPipes API.
TYPE: StreamPipesClient
stream_ids The ids of the data stream to train the model.
TYPE: List[str]
model The model to train. It meant to be a River model/pipeline, but can be every model with a 'learn_one' and 'predict_one' methode.
TYPE: Any
prediction_type The data type of the prediction. Is only needed when you continue to work with the prediction in StreamPipes.
TYPE: str DEFAULT: value
supervised Define if the model is supervised or unsupervised.
TYPE: bool DEFAULT: False
target_label Define the name of the target attribute if the model is supervised.
TYPE: Optional[str] DEFAULT: None
on_start A function to be called when this StreamPipesFunction gets started.
TYPE: Callable[[Any, FunctionContext], None] DEFAULT: lambda self, context: None
on_event A function to be called when this StreamPipesFunction receives an event.
TYPE: Callable[[Any, Dict[str, Any], str], None] DEFAULT: lambda self, event, streamId: None
on_stop A function to be called when this StreamPipesFunction gets stopped.
TYPE: Callable[[Any], None] DEFAULT: lambda self: None
set_learning(learning)","text":"Start or stop the training of the model.
PARAMETER DESCRIPTIONlearning Defines if the training should be continued
TYPE: bool
start()","text":"Registers the function and starts the training.
"},{"location":"reference/function_zoo/river_function/#streampipes.function_zoo.river_function.OnlineML.stop","title":"stop()","text":"Stops the function and ends the training forever.
"},{"location":"reference/function_zoo/river_function/#streampipes.function_zoo.river_function.RiverFunction","title":"RiverFunction(function_definition, model, supervised, target_label, on_start, on_event, on_stop)","text":" Bases: StreamPipesFunction
Implementation of a StreamPipesFunction to enable an easy usage for Online Machine Learning models of the River library.
The function trains the model with the incoming events and publishes the prediction to an output data stream.
PARAMETER DESCRIPTIONfunction_definition The function definition which contains the output stream.
TYPE: FunctionDefinition
model The model to train. It meant to be a River model/pipeline, but can be every model with a 'learn_one' and 'predict_one' method.
TYPE: Any
supervised Define if the model is supervised or unsupervised.
TYPE: bool
target_label Define the name of the target attribute if the model is supervised.
TYPE: Optional[str]
on_start A function to be called when this StreamPipesFunction gets started.
TYPE: Callable[[Any, FunctionContext], None]
on_event A function to be called when this StreamPipesFunction receives an event.
TYPE: Callable[[Any, Dict[str, Any], str], None]
on_stop A function to be called when this StreamPipesFunction gets stopped.
TYPE: Callable[[Any], None]
add_output(stream_id, event)","text":"Send an event via an output data stream to StreamPipes.
PARAMETER DESCRIPTIONstream_id The id of the output data stream
TYPE: str
event The event which should be sent
TYPE: Dict[str, Any]
None"},{"location":"reference/function_zoo/river_function/#streampipes.function_zoo.river_function.RiverFunction.getFunctionId","title":"getFunctionId()","text":"Returns the id of the function.
RETURNS DESCRIPTIONfunction_id Identification object of the StreamPipes function
TYPE: FunctionId
onEvent(event, streamId)","text":"Trains the model with the incoming events and sends the prediction back to StreamPipes.
PARAMETER DESCRIPTIONevent The incoming event that serves as input for the function
TYPE: Dict[str, Any]
streamId Identifier of the corresponding data stream
TYPE: str
None"},{"location":"reference/function_zoo/river_function/#streampipes.function_zoo.river_function.RiverFunction.onServiceStarted","title":"onServiceStarted(context)","text":"Executes the on_start method of the function.
context The functions' context
TYPE: FunctionContext
None"},{"location":"reference/function_zoo/river_function/#streampipes.function_zoo.river_function.RiverFunction.onServiceStopped","title":"onServiceStopped()","text":"Executes the on_stop function.
requiredStreamIds()","text":"Get the ids of the streams needed by the function.
RETURNS DESCRIPTIONstream_ids List of the stream ids
TYPE: List[str]
stop()","text":"Stops the function and disconnects from the output streams.
"},{"location":"reference/functions/function_handler/","title":"Function handler","text":""},{"location":"reference/functions/function_handler/#streampipes.functions.function_handler.FunctionHandler","title":"FunctionHandler(registration, client)","text":"The function handler manages the StreamPipes Functions.
It controls the connection to the brokers, starts the functions, manages the broadcast of the live data and is able to stop the connection to the brokers and functions.
PARAMETER DESCRIPTIONregistration The registration, that contains the StreamPipesFunctions.
TYPE: Registration
client The client to interact with the API.
TYPE: StreamPipesClient
stream_contexts Map of all data stream contexts
TYPE: Dict[str, DataStreamContext]
brokers List of all registered brokers
TYPE: List[Broker]
disconnect()","text":"Disconnects from the brokers and stops all functions.
RETURNS DESCRIPTIONNone"},{"location":"reference/functions/function_handler/#streampipes.functions.function_handler.FunctionHandler.force_stop_functions","title":"force_stop_functions()","text":"Stops the StreamPipesFunctions when the event loop was stopped without stopping the functions.
RETURNS DESCRIPTIONNone WARNS DESCRIPTION UserWarning If there is a running event loop and the functions should be stopped by disconnecting from the broker.
"},{"location":"reference/functions/function_handler/#streampipes.functions.function_handler.FunctionHandler.initializeFunctions","title":"initializeFunctions()","text":"Creates the context for every data stream and starts the event loop to manage the StreamPipes Functions.
RETURNS DESCRIPTIONNone"},{"location":"reference/functions/registration/","title":"Registration","text":""},{"location":"reference/functions/registration/#streampipes.functions.registration.Registration","title":"Registration()","text":"Manages the existing StreamPipesFunctions and registers them.
ATTRIBUTE DESCRIPTIONfunctions List of all registered StreamPipesFunction
TYPE: List[StreamPipesFunction]
getFunctions()","text":"Get all registered functions.
This method exists to be consistent with the Java client.
RETURNS DESCRIPTIONfunctions List of all registered functions.
TYPE: List[StreamPipesFunction]
register(streampipes_function)","text":"Registers a new function.
PARAMETER DESCRIPTIONstreampipes_function The function to register.
TYPE: StreamPipesFunction
self The updated Registration instance
TYPE: Registration
StreamPipesFunction(function_definition=None)","text":" Bases: ABC
Abstract implementation of a StreamPipesFunction.
A StreamPipesFunction allows users to get the data of a StreamPipes data streams easily. It makes it possible to work with the live data in python and enables to use the powerful data analytics libraries there.
PARAMETER DESCRIPTIONfunction_definition The definition of the function that contains metadata about the connected function
TYPE: Optional[FunctionDefinition] DEFAULT: None
output_collectors List of all output collectors which are created based on the provided function definitions.
TYPE: Dict[str, OutputCollector]
add_output(stream_id, event)","text":"Send an event via an output data stream to StreamPipes.
PARAMETER DESCRIPTIONstream_id The id of the output data stream
TYPE: str
event The event which should be sent
TYPE: Dict[str, Any]
None"},{"location":"reference/functions/streampipes_function/#streampipes.functions.streampipes_function.StreamPipesFunction.getFunctionId","title":"getFunctionId()","text":"Returns the id of the function.
RETURNS DESCRIPTIONfunction_id Identification object of the StreamPipes function
TYPE: FunctionId
onEvent(event, streamId) abstractmethod","text":"Is called for every event of a data stream.
PARAMETER DESCRIPTIONevent The received event from the data stream.
TYPE: Dict[str, Any]
streamId The id of the data stream which the event belongs to.
TYPE: str
None"},{"location":"reference/functions/streampipes_function/#streampipes.functions.streampipes_function.StreamPipesFunction.onServiceStarted","title":"onServiceStarted(context) abstractmethod","text":"Is called when the function is started.
PARAMETER DESCRIPTIONcontext The context in which the function is started.
TYPE: FunctionContext
None"},{"location":"reference/functions/streampipes_function/#streampipes.functions.streampipes_function.StreamPipesFunction.onServiceStopped","title":"onServiceStopped() abstractmethod","text":"Is called when the function is stopped.
RETURNS DESCRIPTIONNone"},{"location":"reference/functions/streampipes_function/#streampipes.functions.streampipes_function.StreamPipesFunction.requiredStreamIds","title":"requiredStreamIds()","text":"Get the ids of the streams needed by the function.
RETURNS DESCRIPTIONstream_ids List of the stream ids
TYPE: List[str]
stop()","text":"Stops the function and disconnects from the output streams.
"},{"location":"reference/functions/broker/broker/","title":"Broker","text":""},{"location":"reference/functions/broker/broker/#streampipes.functions.broker.broker.Broker","title":"Broker","text":" Bases: ABC
Abstract implementation of a broker for consumer and publisher.
It contains the basic logic to connect to a data stream.
"},{"location":"reference/functions/broker/broker/#streampipes.functions.broker.broker.Broker.connect","title":"connect(data_stream) async","text":"Connects to the broker running in StreamPipes.
PARAMETER DESCRIPTIONdata_stream Contains the meta information (resources) for a data stream.
TYPE: DataStream
None"},{"location":"reference/functions/broker/broker/#streampipes.functions.broker.broker.Broker.disconnect","title":"disconnect() abstractmethod async","text":"Closes the connection to the server.
RETURNS DESCRIPTIONNone"},{"location":"reference/functions/broker/broker_handler/","title":"Broker handler","text":""},{"location":"reference/functions/broker/broker_handler/#streampipes.functions.broker.broker_handler.SupportedBroker","title":"SupportedBroker","text":" Bases: Enum
Enum for the supported brokers.
"},{"location":"reference/functions/broker/broker_handler/#streampipes.functions.broker.broker_handler.UnsupportedBrokerError","title":"UnsupportedBrokerError(broker_name)","text":" Bases: Exception
Exception if a broker isn't implemented yet.
"},{"location":"reference/functions/broker/broker_handler/#streampipes.functions.broker.broker_handler.get_broker","title":"get_broker(data_stream, is_publisher=False)","text":"Derive the broker for the given data stream.
PARAMETER DESCRIPTIONdata_stream Data stream instance from which the broker is inferred
TYPE: DataStream
broker The corresponding broker instance derived from data stream.
TYPE: Broker
UnsupportedBrokerError Is raised when the given data stream belongs to a broker that is currently not supported by StreamPipes Python.
"},{"location":"reference/functions/broker/broker_handler/#streampipes.functions.broker.broker_handler.get_broker_description","title":"get_broker_description(data_stream)","text":"Derive the decription of the broker for the given data stream.
PARAMETER DESCRIPTIONdata_stream Data stream instance from which the broker is inferred
TYPE: DataStream
broker The corresponding broker description derived from data stream.
TYPE: SupportedBroker
UnsupportedBrokerError Is raised when the given data stream belongs to a broker that is currently not supported by StreamPipes Python.
"},{"location":"reference/functions/broker/consumer/","title":"Consumer","text":""},{"location":"reference/functions/broker/consumer/#streampipes.functions.broker.consumer.Consumer","title":"Consumer","text":" Bases: Broker
Abstract implementation a consumer for a broker.
A consumer allows to subscribe to a data stream.
"},{"location":"reference/functions/broker/consumer/#streampipes.functions.broker.consumer.Consumer.connect","title":"connect(data_stream) async","text":"Connects to the broker running in StreamPipes and creates a subscription.
PARAMETER DESCRIPTIONdata_stream Contains the meta information (resources) for a data stream.
TYPE: DataStream
None"},{"location":"reference/functions/broker/consumer/#streampipes.functions.broker.consumer.Consumer.disconnect","title":"disconnect() abstractmethod async","text":"Closes the connection to the server.
RETURNS DESCRIPTIONNone"},{"location":"reference/functions/broker/consumer/#streampipes.functions.broker.consumer.Consumer.get_message","title":"get_message() abstractmethod","text":"Get the published messages of the subscription.
RETURNS DESCRIPTIONiterator An async iterator for the messages.
TYPE: AsyncIterator
OutputCollector(data_stream)","text":"Collector for output events. The events are published to an output data stream. Therefore, the output collector establishes a connection to the broker.
PARAMETER DESCRIPTIONdata_stream The output data stream that will receive the events.
TYPE: DataStream
publisher The publisher instance that sends the data to StreamPipes
TYPE: Publisher
collect(event)","text":"Publishes an event to the output stream.
PARAMETER DESCRIPTIONevent The event to be published.
TYPE: Dict[str, Any]
None"},{"location":"reference/functions/broker/output_collector/#streampipes.functions.broker.output_collector.OutputCollector.disconnect","title":"disconnect()","text":"Disconnects the broker of the output collector.
RETURNS DESCRIPTIONNone"},{"location":"reference/functions/broker/publisher/","title":"Publisher","text":""},{"location":"reference/functions/broker/publisher/#streampipes.functions.broker.publisher.Publisher","title":"Publisher","text":" Bases: Broker
Abstract implementation of a publisher for a broker.
A publisher allows to publish events to a data stream.
"},{"location":"reference/functions/broker/publisher/#streampipes.functions.broker.publisher.Publisher.connect","title":"connect(data_stream) async","text":"Connects to the broker running in StreamPipes.
PARAMETER DESCRIPTIONdata_stream Contains the meta information (resources) for a data stream.
TYPE: DataStream
None"},{"location":"reference/functions/broker/publisher/#streampipes.functions.broker.publisher.Publisher.disconnect","title":"disconnect() abstractmethod async","text":"Closes the connection to the server.
RETURNS DESCRIPTIONNone"},{"location":"reference/functions/broker/publisher/#streampipes.functions.broker.publisher.Publisher.publish_event","title":"publish_event(event) abstractmethod async","text":"Publish an event to a connected data stream.
PARAMETER DESCRIPTIONevent The event to be published.
TYPE: Dict[str, Any]
None"},{"location":"reference/functions/broker/kafka/kafka_consumer/","title":"Kafka consumer","text":""},{"location":"reference/functions/broker/kafka/kafka_consumer/#streampipes.functions.broker.kafka.kafka_consumer.KafkaConsumer","title":"KafkaConsumer","text":" Bases: Consumer
Implementation of a consumer for Kafka
"},{"location":"reference/functions/broker/kafka/kafka_consumer/#streampipes.functions.broker.kafka.kafka_consumer.KafkaConsumer.connect","title":"connect(data_stream) async","text":"Connects to the broker running in StreamPipes and creates a subscription.
PARAMETER DESCRIPTIONdata_stream Contains the meta information (resources) for a data stream.
TYPE: DataStream
None"},{"location":"reference/functions/broker/kafka/kafka_consumer/#streampipes.functions.broker.kafka.kafka_consumer.KafkaConsumer.disconnect","title":"disconnect() async","text":"Closes the connection to the server.
RETURNS DESCRIPTIONNone"},{"location":"reference/functions/broker/kafka/kafka_consumer/#streampipes.functions.broker.kafka.kafka_consumer.KafkaConsumer.get_message","title":"get_message()","text":"Get the published messages of the subscription.
RETURNS DESCRIPTIONiterator An async iterator for the messages.
TYPE: AsyncIterator
KafkaMessage(data)","text":"An internal representation of a Kafka message
PARAMETER DESCRIPTIONdata The received Kafka message as byte array
"},{"location":"reference/functions/broker/kafka/kafka_message_fetcher/#streampipes.functions.broker.kafka.kafka_message_fetcher.KafkaMessageFetcher","title":"
KafkaMessageFetcher(consumer)","text":"Fetches the next message from Kafka
PARAMETER DESCRIPTIONconsumer The Kafka consumer
TYPE: Consumer
KafkaPublisher","text":" Bases: Publisher
Implementation of a publisher for Kafka
"},{"location":"reference/functions/broker/kafka/kafka_publisher/#streampipes.functions.broker.kafka.kafka_publisher.KafkaPublisher.connect","title":"connect(data_stream) async","text":"Connects to the broker running in StreamPipes.
PARAMETER DESCRIPTIONdata_stream Contains the meta information (resources) for a data stream.
TYPE: DataStream
None"},{"location":"reference/functions/broker/kafka/kafka_publisher/#streampipes.functions.broker.kafka.kafka_publisher.KafkaPublisher.disconnect","title":"disconnect() async","text":"Closes the connection to the server.
RETURNS DESCRIPTIONNone"},{"location":"reference/functions/broker/kafka/kafka_publisher/#streampipes.functions.broker.kafka.kafka_publisher.KafkaPublisher.publish_event","title":"publish_event(event) async","text":"Publish an event to a connected data stream.
PARAMETER DESCRIPTIONevent The event to be published.
TYPE: Dict[str, Any]
None"},{"location":"reference/functions/broker/nats/nats_consumer/","title":"Nats consumer","text":""},{"location":"reference/functions/broker/nats/nats_consumer/#streampipes.functions.broker.nats.nats_consumer.NatsConsumer","title":"NatsConsumer","text":" Bases: Consumer
Implementation of a consumer for NATS
"},{"location":"reference/functions/broker/nats/nats_consumer/#streampipes.functions.broker.nats.nats_consumer.NatsConsumer.connect","title":"connect(data_stream) async","text":"Connects to the broker running in StreamPipes and creates a subscription.
PARAMETER DESCRIPTIONdata_stream Contains the meta information (resources) for a data stream.
TYPE: DataStream
None"},{"location":"reference/functions/broker/nats/nats_consumer/#streampipes.functions.broker.nats.nats_consumer.NatsConsumer.disconnect","title":"disconnect() async","text":"Closes the connection to the server.
RETURNS DESCRIPTIONNone"},{"location":"reference/functions/broker/nats/nats_consumer/#streampipes.functions.broker.nats.nats_consumer.NatsConsumer.get_message","title":"get_message()","text":"Get the published messages of the subscription.
RETURNS DESCRIPTIONmessage_iterator An async iterator for the messages.
TYPE: AsyncIterator
NatsPublisher","text":" Bases: Publisher
Implementation of a publisher for NATS
"},{"location":"reference/functions/broker/nats/nats_publisher/#streampipes.functions.broker.nats.nats_publisher.NatsPublisher.connect","title":"connect(data_stream) async","text":"Connects to the broker running in StreamPipes.
PARAMETER DESCRIPTIONdata_stream Contains the meta information (resources) for a data stream.
TYPE: DataStream
None"},{"location":"reference/functions/broker/nats/nats_publisher/#streampipes.functions.broker.nats.nats_publisher.NatsPublisher.disconnect","title":"disconnect() async","text":"Closes the connection to the server.
RETURNS DESCRIPTIONNone"},{"location":"reference/functions/broker/nats/nats_publisher/#streampipes.functions.broker.nats.nats_publisher.NatsPublisher.publish_event","title":"publish_event(event) async","text":"Publish an event to a connected data stream.
PARAMETER DESCRIPTIONevent The event to be published.
TYPE: Dict[str, Any]
None"},{"location":"reference/functions/utils/async_iter_handler/","title":"Async iter handler","text":""},{"location":"reference/functions/utils/async_iter_handler/#streampipes.functions.utils.async_iter_handler.AsyncIterHandler","title":"AsyncIterHandler","text":"Handles asynchronous iterators to get every message after another in parallel.
"},{"location":"reference/functions/utils/async_iter_handler/#streampipes.functions.utils.async_iter_handler.AsyncIterHandler.anext","title":"anext(stream_id, message) async staticmethod","text":"Gets the next message from an AsyncIterator.
PARAMETER DESCRIPTIONstream_id The id of the data stream which the message belongs to.
TYPE: str
message An asynchronous iterator that contains the messages.
TYPE: AsyncIterator
result Tuple of the stream id und next message or (\"stop\", None) if no message is left.
TYPE: Tuple[str, Optional[Any]]
combine_async_messages(messages) async staticmethod","text":"Continuously gets the next published message from multiple AsyncIterators in parallel.
PARAMETER DESCRIPTIONmessages A dictionary with an asynchronous iterator for every stream id.
TYPE: Dict[str, AsyncIterator]
message Description of the anonymous integer return value.
TYPE:: Tuple[str, Any]
DataStreamContext(functions, schema, broker)","text":"Container for the context of a data stream.
PARAMETER DESCRIPTIONfunctions StreamPipes Functions which require the data of this data stream.
TYPE: List[StreamPipesFunction]
schema The schema of this data stream.
TYPE: DataStream
broker The consumer to connect to this data stream.
TYPE: Consumer
add_function(function)","text":"Adds a new StreamPipes Function.
PARAMETER DESCRIPTIONfunction StreamPipesFunction which requires this data stream.
TYPE: StreamPipesFunction
None"},{"location":"reference/functions/utils/data_stream_generator/","title":"Data stream generator","text":""},{"location":"reference/functions/utils/data_stream_generator/#streampipes.functions.utils.data_stream_generator.RuntimeType","title":"RuntimeType","text":" Bases: Enum
Runtime type names for the attributes of a data stream.
ATTRIBUTE DESCRIPTIONSTRING
BOOLEAN
DOUBLE
FLOAT
INTEGER
LONG "},{"location":"reference/functions/utils/data_stream_generator/#streampipes.functions.utils.data_stream_generator.create_data_stream","title":"
create_data_stream(name, attributes, stream_id=None, broker=SupportedBroker.NATS)","text":"Creates a data stream
PARAMETER DESCRIPTIONname Name of the data stream to be shown at the UI.
TYPE: str
attributes Name and types of the attributes.
TYPE: Dict[str, str]
stream_id The id of this data stream.
TYPE: str DEFAULT: None
data_stream The created data stream
TYPE: DataStream
FunctionContext(function_id, schema, client, streams)","text":"Container for the context of a StreamPipesFunction.
PARAMETER DESCRIPTIONfunction_id The id of this function.
TYPE: str
schema A dictionary which contains the schema of a data stream for each stream id.
TYPE: Dict[str, DataStream]
client The client to interact with the API.
TYPE: StreamPipesClient
streams The ids of the streams needed by this function.
TYPE: List[str]
add_data_stream_schema(stream_id, data_stream)","text":"Adds a new data stream for a new stream id.
PARAMETER DESCRIPTIONstream_id The id of the data stream.
TYPE: str
data_stream The schema of the data stream.
TYPE: DataStream
None"},{"location":"reference/model/common/","title":"Common","text":"Classes of the StreamPipes data model that are commonly shared.
"},{"location":"reference/model/common/#streampipes.model.common.ApplicationLink","title":"ApplicationLink","text":" Bases: BasicModel
Data model of an ApplicationLink in compliance with the StreamPipes Backend.
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/common/#streampipes.model.common.BaseElement","title":"BaseElement","text":" Bases: BasicModel
Structure of a basic element in the StreamPipes Backend.
"},{"location":"reference/model/common/#streampipes.model.common.BaseElement.Config","title":"Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/common/#streampipes.model.common.BasicModel","title":"BasicModel","text":" Bases: BaseModel
Basic model class used for the whole Python StreamPipes data model.
"},{"location":"reference/model/common/#streampipes.model.common.BasicModel.Config","title":"Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/common/#streampipes.model.common.EventGrounding","title":"EventGrounding","text":" Bases: BasicModel
Data model of an EventGrounding in compliance to with StreamPipes Backend.
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/common/#streampipes.model.common.EventProperty","title":"EventProperty","text":" Bases: BasicModel
Data model of an EventProperty in compliance with the StreamPipes Backend.
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/common/#streampipes.model.common.EventSchema","title":"EventSchema","text":" Bases: BasicModel
Data model of an EventSchema in compliance with the StreamPipes Backend.
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/common/#streampipes.model.common.MeasurementCapability","title":"MeasurementCapability","text":" Bases: BasicModel
Data model of a MeasurementCapability in compliance with the StreamPipes Backend.
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/common/#streampipes.model.common.MeasurementObject","title":"MeasurementObject","text":" Bases: BasicModel
Data model of a MeasurementObject in compliance with the StreamPipes Backend.
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/common/#streampipes.model.common.TopicDefinition","title":"TopicDefinition","text":" Bases: BasicModel
Data model of a TopicDefinition in compliance with the StreamPipes Backend.
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/common/#streampipes.model.common.TransportFormat","title":"TransportFormat","text":" Bases: BasicModel
Data model of a TransportFormat in compliance with the StreamPipes Backend.
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/common/#streampipes.model.common.TransportProtocol","title":"TransportProtocol","text":" Bases: BasicModel
Data model of a TransportProtocol in compliance with the StreamPipes Backend.
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/common/#streampipes.model.common.ValueSpecification","title":"ValueSpecification","text":" Bases: BasicModel
Data model of an ValueSpecification in compliance with the StreamPipes Backend.
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/common/#streampipes.model.common.random_letters","title":"random_letters(n)","text":"Generates a string consisting of random letters.
PARAMETER DESCRIPTIONn number of letters
TYPE: int
rand_str String consisting of n random letters
TYPE: str
Implementation of a resource container for the data lake measures endpoint.
"},{"location":"reference/model/container/data_lake_measures/#streampipes.model.container.data_lake_measures.DataLakeMeasures","title":"DataLakeMeasures(resources)","text":" Bases: ResourceContainer
Implementation of the resource container for the data lake measures endpoint.
This resource container is a collection of data lake measures returned by the StreamPipes API. It is capable of parsing the response content directly into a list of queried DataLakeMeasure. Furthermore, the resource container makes them accessible in a pythonic manner.
from_json(json_string) classmethod","text":"Creates a ResourceContainer from the given JSON string.
json_string The JSON string returned from the StreamPipes API.
TYPE: str
container Instance of the container derived from the JSON definition
TYPE: ResourceContainer
StreamPipesDataModelError If a resource cannot be mapped to the corresponding Python data model.
StreamPipesResourceContainerJSONError If JSON response cannot be parsed to a ResourceContainer.
to_dicts(use_source_names=False)","text":"Returns the contained resources as list of dictionaries.
PARAMETER DESCRIPTIONuse_source_names Determines whether the field names are named in Python style (=False) or as originally named by StreamPipes (=True).
TYPE: bool DEFAULT: False
dictionary_list List of resources in dictionary representation. If use_source_names equals True the keys are named as in the StreamPipes backend.
TYPE: List[Dict[str, Any]]
to_json()","text":"Returns the resource container in the StreamPipes JSON representation.
RETURNS DESCRIPTION json_string: str JSON representation of the resource container where key names are equal to keys used in the StreamPipes backend
"},{"location":"reference/model/container/data_lake_measures/#streampipes.model.container.data_lake_measures.DataLakeMeasures.to_pandas","title":"to_pandas()","text":"Returns the resource container in representation of a Pandas Dataframe.
RETURNS DESCRIPTIONresource_container_df Representation of the resource container as pandas DataFrame
TYPE: DataFrame
Implementation of a resource container for the data streams endpoint.
"},{"location":"reference/model/container/data_streams/#streampipes.model.container.data_streams.DataStreams","title":"DataStreams(resources)","text":" Bases: ResourceContainer
Implementation of the resource container for the data stream endpoint.
This resource container is a collection of data streams returned by the StreamPipes API. It is capable of parsing the response content directly into a list of queried DataStream. Furthermore, the resource container makes them accessible in a pythonic manner.
from_json(json_string) classmethod","text":"Creates a ResourceContainer from the given JSON string.
json_string The JSON string returned from the StreamPipes API.
TYPE: str
container Instance of the container derived from the JSON definition
TYPE: ResourceContainer
StreamPipesDataModelError If a resource cannot be mapped to the corresponding Python data model.
StreamPipesResourceContainerJSONError If JSON response cannot be parsed to a ResourceContainer.
to_dicts(use_source_names=False)","text":"Returns the contained resources as list of dictionaries.
PARAMETER DESCRIPTIONuse_source_names Determines whether the field names are named in Python style (=False) or as originally named by StreamPipes (=True).
TYPE: bool DEFAULT: False
dictionary_list List of resources in dictionary representation. If use_source_names equals True the keys are named as in the StreamPipes backend.
TYPE: List[Dict[str, Any]]
to_json()","text":"Returns the resource container in the StreamPipes JSON representation.
RETURNS DESCRIPTION json_string: str JSON representation of the resource container where key names are equal to keys used in the StreamPipes backend
"},{"location":"reference/model/container/data_streams/#streampipes.model.container.data_streams.DataStreams.to_pandas","title":"to_pandas()","text":"Returns the resource container in representation of a Pandas Dataframe.
RETURNS DESCRIPTIONresource_container_df Representation of the resource container as pandas DataFrame
TYPE: DataFrame
General and abstract implementation for a resource container.
A resource container is a collection of resources returned by the StreamPipes API. It is capable of parsing the response content directly into a list of queried resources. Furthermore, the resource container makes them accessible in a pythonic manner.
"},{"location":"reference/model/container/resource_container/#streampipes.model.container.resource_container.ResourceContainer","title":"ResourceContainer(resources)","text":" Bases: ABC
General and abstract implementation for a resource container.
A resource container is a collection of resources returned by the StreamPipes API. It is capable of parsing the response content directly into a list of queried resources. Furthermore, the resource container makes them accessible in a pythonic manner.
PARAMETER DESCRIPTIONresources A list of resources to be contained in the ResourceContainer.
TYPE: List[Resource]
from_json(json_string) classmethod","text":"Creates a ResourceContainer from the given JSON string.
json_string The JSON string returned from the StreamPipes API.
TYPE: str
container Instance of the container derived from the JSON definition
TYPE: ResourceContainer
StreamPipesDataModelError If a resource cannot be mapped to the corresponding Python data model.
StreamPipesResourceContainerJSONError If JSON response cannot be parsed to a ResourceContainer.
to_dicts(use_source_names=False)","text":"Returns the contained resources as list of dictionaries.
PARAMETER DESCRIPTIONuse_source_names Determines whether the field names are named in Python style (=False) or as originally named by StreamPipes (=True).
TYPE: bool DEFAULT: False
dictionary_list List of resources in dictionary representation. If use_source_names equals True the keys are named as in the StreamPipes backend.
TYPE: List[Dict[str, Any]]
to_json()","text":"Returns the resource container in the StreamPipes JSON representation.
RETURNS DESCRIPTION json_string: str JSON representation of the resource container where key names are equal to keys used in the StreamPipes backend
"},{"location":"reference/model/container/resource_container/#streampipes.model.container.resource_container.ResourceContainer.to_pandas","title":"to_pandas()","text":"Returns the resource container in representation of a Pandas Dataframe.
RETURNS DESCRIPTIONresource_container_df Representation of the resource container as pandas DataFrame
TYPE: DataFrame
StreamPipesDataModelError(validation_error)","text":" Bases: Exception
A custom exception to be raised when a validation error occurs during the parsing of StreamPipes API responses.
PARAMETER DESCRIPTIONvalidation_error The validation error thrown by Pydantic during parsing.
TYPE: ValidationError
StreamPipesResourceContainerJSONError(container_name, json_string)","text":" Bases: Exception
A custom exception to be raised when the returned JSON string does not suit to the structure of resource container.
PARAMETER DESCRIPTIONcontainer_name The class name of the resource container where the invalid data structure was detected.
TYPE: str
json_string The JSON string that has been tried to parse.
TYPE: str
Implementation of a resource container for the versions endpoint.
"},{"location":"reference/model/container/versions/#streampipes.model.container.versions.Versions","title":"Versions(resources)","text":" Bases: ResourceContainer
Implementation of the resource container for the versions endpoint.
This resource container is a collection of versions returned by the StreamPipes API. It is capable of parsing the response content directly into a list of queried Version. Furthermore, the resource container makes them accessible in a pythonic manner.
resources A list of resources (Version) to be contained in the ResourceContainer.
TYPE: List[Resource]
from_json(json_string) classmethod","text":"Creates a ResourceContainer from the given JSON string.
json_string The JSON string returned from the StreamPipes API.
TYPE: str
container Instance of the container derived from the JSON definition
TYPE: ResourceContainer
StreamPipesDataModelError If a resource cannot be mapped to the corresponding Python data model.
StreamPipesResourceContainerJSONError If JSON response cannot be parsed to a ResourceContainer.
to_dicts(use_source_names=False)","text":"Returns the contained resources as list of dictionaries.
PARAMETER DESCRIPTIONuse_source_names Determines whether the field names are named in Python style (=False) or as originally named by StreamPipes (=True).
TYPE: bool DEFAULT: False
dictionary_list List of resources in dictionary representation. If use_source_names equals True the keys are named as in the StreamPipes backend.
TYPE: List[Dict[str, Any]]
to_json()","text":"Returns the resource container in the StreamPipes JSON representation.
RETURNS DESCRIPTION json_string: str JSON representation of the resource container where key names are equal to keys used in the StreamPipes backend
"},{"location":"reference/model/container/versions/#streampipes.model.container.versions.Versions.to_pandas","title":"to_pandas()","text":"Returns the resource container in representation of a Pandas Dataframe.
RETURNS DESCRIPTIONresource_container_df Representation of the resource container as pandas DataFrame
TYPE: DataFrame
DataLakeMeasure","text":" Bases: Resource
Implementation of a resource for data lake measures.
This resource defines the data model used by resource container (model.container.DataLakeMeasures). It inherits from Pydantic's BaseModel to get all its superpowers, which are used to parse, validate the API response, and to easily switch between the Python representation (both serialized and deserialized) and Java representation (serialized only).
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/resource/data_lake_measure/#streampipes.model.resource.data_lake_measure.DataLakeMeasure.convert_to_pandas_representation","title":"convert_to_pandas_representation()","text":"Returns the dictionary representation of a data lake measure to be used when creating a pandas Dataframe.
It excludes the following fields: element_id, event_schema, schema_version. Instead of the whole event schema the number of event properties contained is returned with the column name num_event_properties.
pandas_repr Pandas representation of the resource as a dictionary, which is then used by the respource container to create a data frame from a collection of resources.
TYPE: Dict[str, Any]
to_dict(use_source_names=True)","text":"Returns the resource in dictionary representation.
PARAMETER DESCRIPTIONuse_source_names Indicates if the dictionary keys are in python representation or equally named to the StreamPipes backend
DEFAULT: True
resource The resource as dictionary representation
TYPE: Dict[str, Any]
DataSeries","text":" Bases: Resource
Implementation of a resource for data series. This resource defines the data model used by its resource container(model.container.DataLakeMeasures). It inherits from Pydantic's BaseModel to get all its superpowers, which are used to parse, validate the API response and to easily switch between the Python representation (both serialized and deserialized) and Java representation (serialized only).
This class will only exist temporarily in it its current appearance since\nthere are some inconsistencies in the StreamPipes API.\nConfig","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/resource/data_series/#streampipes.model.resource.data_series.DataSeries.convert_to_pandas_representation","title":"convert_to_pandas_representation()","text":"Returns the dictionary representation of a data lake series to be used when creating a pandas Dataframe.
It contains only the \"header rows\" (the column names) and \"rows\" that contain the actual data.
RETURNS DESCRIPTIONpandas_repr Dictionary with the keys headers and rows
TYPE: dict[str, Any]
from_json(json_string) classmethod","text":"Creates an instance of DataSeries from a given JSON string.
This method is used by the resource container to parse the JSON response of the StreamPipes API. Currently, it only supports data lake series that consist of exactly one series of data.
PARAMETER DESCRIPTIONjson_string The JSON string the data lake series should be created on.
TYPE: str
DataSeries Instance of DataSeries that is created based on the given JSON string.
StreamPipesUnsupportedDataLakeSeries If the data lake series returned by the StreamPipes API cannot be parsed with the current version of the Python client.
"},{"location":"reference/model/resource/data_series/#streampipes.model.resource.data_series.DataSeries.to_dict","title":"to_dict(use_source_names=True)","text":"Returns the resource in dictionary representation.
PARAMETER DESCRIPTIONuse_source_names Indicates if the dictionary keys are in python representation or equally named to the StreamPipes backend
DEFAULT: True
resource The resource as dictionary representation
TYPE: Dict[str, Any]
to_pandas()","text":"Returns the data lake series in representation of a Pandas Dataframe.
RETURNS DESCRIPTIONpd The data lake series in form of a pandas dataframe
TYPE: DataFrame
DataStream(**kwargs)","text":" Bases: Resource
Implementation of a resource for data streams.
This resource defines the data model used by resource container (model.container.DataStreams). It inherits from Pydantic's BaseModel to get all its superpowers, which are used to parse, validate the API response and to easily switch between the Python representation (both serialized and deserialized) and Java representation (serialized only).
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/resource/data_stream/#streampipes.model.resource.data_stream.DataStream.convert_to_pandas_representation","title":"convert_to_pandas_representation()","text":"Returns the dictionary representation of a data stream to be used when creating a pandas Dataframe.
RETURNS DESCRIPTIONpandas_repr Pandas representation of the resource as a dictionary, which is then used by the respource container to create a data frame from a collection of resources.
TYPE: Dict[str, Any]
to_dict(use_source_names=True)","text":"Returns the resource in dictionary representation.
PARAMETER DESCRIPTIONuse_source_names Indicates if the dictionary keys are in python representation or equally named to the StreamPipes backend
DEFAULT: True
resource The resource as dictionary representation
TYPE: Dict[str, Any]
StreamPipesUnsupportedDataSeries(reason=None)","text":" Bases: Exception
Exception to be raised when the returned data lake series cannot be parsed with the current implementation of the resource.
"},{"location":"reference/model/resource/function_definition/","title":"Function definition","text":""},{"location":"reference/model/resource/function_definition/#streampipes.model.resource.function_definition.FunctionDefinition","title":"FunctionDefinition","text":" Bases: Resource
Configuration for a StreamPipes Function.
This class maps to the FunctionDefinition class in the StreamPipes model. It contains all metadata that are required to register a function at the StreamPipes backend.
consumed_streams List of data streams the function is consuming from
TYPE: List[str]
function_id identifier object of a StreamPipes function
TYPE: FunctionId
output_data_streams Map off all output data streams added to the function definition
TYPE: Dict[str, DataStream]
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/resource/function_definition/#streampipes.model.resource.function_definition.FunctionDefinition.add_output_data_stream","title":"add_output_data_stream(data_stream)","text":"Adds an output data stream to the function which makes it possible to write data back to StreamPipes.
PARAMETER DESCRIPTIONdata_stream The schema of the output data stream.
TYPE: DataStream
self Instance of the function definition that is extended by the provided DataStream
TYPE: FunctionDefinition
convert_to_pandas_representation()","text":"Returns the dictionary representation of a function definition to be used when creating a pandas Dataframe.
RETURNS DESCRIPTIONpandas_repr Pandas representation of the resource as a dictionary, which is then used by the respource container to create a data frame from a collection of resources.
TYPE: Dict[str, Any]
get_output_data_streams()","text":"Get the output data streams of the function.
RETURNS DESCRIPTIONoutput_streams Dictionary with every known stream id and the related output stream.
TYPE: Dict[str, DataStream]
get_output_stream_ids()","text":"Get the stream ids of the output data streams.
RETURNS DESCRIPTIONoutput_stream_ids List of all stream ids
TYPE: List[str]
to_dict(use_source_names=True)","text":"Returns the resource in dictionary representation.
PARAMETER DESCRIPTIONuse_source_names Indicates if the dictionary keys are in python representation or equally named to the StreamPipes backend
DEFAULT: True
resource The resource as dictionary representation
TYPE: Dict[str, Any]
FunctionId","text":" Bases: BasicModel
Identification object for a StreamPipes function.
Maps to the FunctionId class defined in the StreamPipes model.
id unique identifier of the function instance
TYPE: str
version version of the corresponding function
TYPE: int
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/resource/query_result/","title":"Query result","text":""},{"location":"reference/model/resource/query_result/#streampipes.model.resource.query_result.QueryResult","title":"QueryResult","text":" Bases: Resource
Implementation of a resource for query result. This resource defines the data model used by its resource container(model.container.DataLakeMeasures). It inherits from Pydantic's BaseModel to get all its superpowers, which are used to parse, validate the API response and to easily switch between the Python representation (both serialized and deserialized) and Java representation (serialized only).
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/resource/query_result/#streampipes.model.resource.query_result.QueryResult.convert_to_pandas_representation","title":"convert_to_pandas_representation()","text":"Returns the dictionary representation of a data lake series to be used when creating a pandas Dataframe.
It contains only the \"header rows\" (the column names) and \"rows\" that contain the actual data.
RETURNS DESCRIPTIONdict Dictionary with the keys headers and rows
StreamPipesUnsupportedDataLakeSeries If the query result returned by the StreamPipes API cannot be converted to the pandas representation
"},{"location":"reference/model/resource/query_result/#streampipes.model.resource.query_result.QueryResult.to_dict","title":"to_dict(use_source_names=True)","text":"Returns the resource in dictionary representation.
PARAMETER DESCRIPTIONuse_source_names Indicates if the dictionary keys are in python representation or equally named to the StreamPipes backend
DEFAULT: True
resource The resource as dictionary representation
TYPE: Dict[str, Any]
to_pandas()","text":"Returns the data lake series in representation of a Pandas Dataframe.
RETURNS DESCRIPTIONdf Pandas df containing the query result
TYPE: DataFrame
General and abstract implementation for a resource.
A resource defines the data model that is used by a resource container (model.container.resourceContainer).
Resource","text":" Bases: ABC, BasicModel
General and abstract implementation for a resource.
A resource defines the data model used by a resource container (model.container.resourceContainer). It inherits from Pydantic's BaseModel to get all its superpowers, which are used to parse, validate the API response and to easily switch between the Python representation (both serialized and deserialized) and Java representation (serialized only).
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/resource/resource/#streampipes.model.resource.resource.Resource.convert_to_pandas_representation","title":"convert_to_pandas_representation() abstractmethod","text":"Returns a dictionary representation to be used when creating a pandas Dataframe.
RETURNS DESCRIPTIONpandas_repr Pandas representation of the resource as a dictionary, which is then used by the respource container to create a data frame from a collection of resources.
TYPE: Dict[str, Any]
to_dict(use_source_names=True)","text":"Returns the resource in dictionary representation.
PARAMETER DESCRIPTIONuse_source_names Indicates if the dictionary keys are in python representation or equally named to the StreamPipes backend
DEFAULT: True
resource The resource as dictionary representation
TYPE: Dict[str, Any]
Version","text":" Bases: Resource
Metadata about the version of the connected StreamPipes server.
ATTRIBUTE DESCRIPTIONbackend_version version of the StreamPipes backend the client is connected to
TYPE: str
Config","text":"Configuration class for Pydantic. Defines alias generator to convert field names from camelCase (API) to snake_case (Python codebase).
"},{"location":"reference/model/resource/version/#streampipes.model.resource.version.Version.convert_to_pandas_representation","title":"convert_to_pandas_representation()","text":"Returns the dictionary representation of the version metadata to be used when creating a pandas Dataframe.
"},{"location":"reference/model/resource/version/#streampipes.model.resource.version.Version.to_dict","title":"to_dict(use_source_names=True)","text":"Returns the resource in dictionary representation.
PARAMETER DESCRIPTIONuse_source_names Indicates if the dictionary keys are in python representation or equally named to the StreamPipes backend
DEFAULT: True
resource The resource as dictionary representation
TYPE: Dict[str, Any]
validate_backend_version(backend_version)","text":"Validates the backend version of the StreamPipes. Sets 'development' if none is returned since this the behavior of StreamPipes backend running in development mode.
"},{"location":"tutorials/1-introduction-to-streampipes-python-client/","title":"Introduction to StreamPipes Python","text":"In\u00a0[\u00a0]: Copied!%pip install streampipes\n%pip install streampipes
If you want to have the current development state you can also execute:
In\u00a0[\u00a0]: Copied!%pip install git+https://github.com/apache/streampipes.git#subdirectory=streampipes-client-python\n%pip install git+https://github.com/apache/streampipes.git#subdirectory=streampipes-client-python
The corresponding documentation can be found here.
In\u00a0[\u00a0]: Copied!from streampipes.client import StreamPipesClient\nfrom streampipes.client.config import StreamPipesClientConfig\nfrom streampipes.client.credential_provider import StreamPipesApiKeyCredentials\nfrom streampipes.client import StreamPipesClient from streampipes.client.config import StreamPipesClientConfig from streampipes.client.credential_provider import StreamPipesApiKeyCredentials In\u00a0[\u00a0]: Copied!
config = StreamPipesClientConfig(\n credential_provider=StreamPipesApiKeyCredentials(\n username=\"test@streampipes.apache.org\",\n api_key=\"API-KEY\",\n ),\n host_address=\"localhost\",\n https_disabled=True,\n port=80\n)\nconfig = StreamPipesClientConfig( credential_provider=StreamPipesApiKeyCredentials( username=\"test@streampipes.apache.org\", api_key=\"API-KEY\", ), host_address=\"localhost\", https_disabled=True, port=80 )
Please be aware that connecting to StreamPipes via a https connection is currently not supported by the Python client.
Providing secrets like the api_key as plaintext in the source code is an anti-pattern. This is why the StreamPipes client also supports passing the required secrets as environment variables. To do so, you must initialize the credential provider like the following:
StreamPipesApiKeyCredentials()\nStreamPipesApiKeyCredentials()
To ensure that the above code works, you must set the environment variables as expected. This can be done as follows:
In\u00a0[\u00a0]: Copied!import os\nos.environ[\"SP_USERNAME\"] = \"admin@streampipes.apache.org\"\nos.environ[\"SP_API_KEY\"] = \"XXX\"\nimport os os.environ[\"SP_USERNAME\"] = \"admin@streampipes.apache.org\" os.environ[\"SP_API_KEY\"] = \"XXX\"
Having the config ready, we can now initialize the actual client.
client = StreamPipesClient(client_config=config)\nclient = StreamPipesClient(client_config=config)
That's already it. You can check if everything works out by using the following command:
In\u00a0[6]: Copied!client.describe()\nclient.describe()
2023-02-24 17:05:49,398 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:167] [_make_request] - Successfully retrieved all resources.\n2023-02-24 17:05:49,457 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:167] [_make_request] - Successfully retrieved all resources.\n\nHi there!\nYou are connected to a StreamPipes instance running at http://localhost:80.\nThe following StreamPipes resources are available with this client:\n1x DataLakeMeasures\n1x DataStreams\n
This prints you a short textual description of the connected StreamPipes instance to the console.
The created client instance serves as the central point of interaction with StreamPipes. You can invoke a variety of commands directly on this object.
Are you curious now how you actually can get data out of StreamPipes and make use of it with Python? Then check out the next tutorial on extracting Data from the StreamPipes data lake.
Thanks for reading this introductory tutorial. We hope you like it and would love to receive some feedback from you. Just go to our GitHub discussion page and let us know your impression. We'll read and react to them all, we promise!
"},{"location":"tutorials/1-introduction-to-streampipes-python-client/#Introduction-to-StreamPipes-Python","title":"Introduction to StreamPipes Python\u00b6","text":""},{"location":"tutorials/1-introduction-to-streampipes-python-client/#Why-there-is-an-extra-Python-library-for-StreamPipes?","title":"Why there is an extra Python library for StreamPipes?\u00b6","text":"Apache StreamPipes aims to enable non-technical users to connect and analyze IoT data streams. To achieve this, it provides an easy-to-use and convenient user interface that allows one to connect to an IoT data source and create some visual graphs within a few minutes. While this is the primary use case for Apache StreamPipes, it also offers significant value to those interested in data analysis or data science with IoT data, without the need to handle the complexities of extracting data from devices in a suitable format. In this scenario, StreamPipes helps you connect to your data source and extract the data for you. You then can make the data available outside StreamPipes by writing it into an external source, such as a database, Kafka, etc. While this requires another component, you can also extract your data directly from StreamPipes programmatically using the StreamPipes API. For convenience, we also provide you with a StreamPipes client both available for Java and Python. Specifically with StreamPipes Python, we want to address the amazing data analytics and data science community in Python and benefit from the great universe of Python libraries out there.
"},{"location":"tutorials/1-introduction-to-streampipes-python-client/#How-to-install-StreamPipes-Python?","title":"How to install StreamPipes Python?\u00b6","text":"Simply use the following pip command:
In case you want to reproduce the first two tutorials exactly on your end, you need to create a simple pipeline in StreamPipes like demonstrated below.
"},{"location":"tutorials/1-introduction-to-streampipes-python-client/#How-to-configure-the-Python-client","title":"How to configure the Python client\u00b6","text":"In order to access the resources available in StreamPipes, one must be able to authenticate against the backend. For this purpose, the client so far only supports the authentication via an API token that can be generated via the StreamPipes UI, as you can see below.
Having generated the API token, one can directly start initializing a client instance as follows:
"},{"location":"tutorials/2-extracting-data-from-the-streampipes-data-lake/","title":"Extracting Data from the StreamPipes data lake","text":"In\u00a0[1]: Copied!from streampipes.client import StreamPipesClient\nfrom streampipes.client.config import StreamPipesClientConfig\nfrom streampipes.client.credential_provider import StreamPipesApiKeyCredentials\nfrom streampipes.client import StreamPipesClient from streampipes.client.config import StreamPipesClientConfig from streampipes.client.credential_provider import StreamPipesApiKeyCredentials In\u00a0[\u00a0]: Copied!
# if you want all necessary dependencies required for this tutorial to be installed,\n# you can simply execute the following command\n%pip install matplotlib streampipes\n# if you want all necessary dependencies required for this tutorial to be installed, # you can simply execute the following command %pip install matplotlib streampipes In\u00a0[2]: Copied!
import os\nos.environ[\"SP_USERNAME\"] = \"admin@streampipes.apache.org\"\nos.environ[\"SP_API_KEY\"] = \"XXX\"\nimport os os.environ[\"SP_USERNAME\"] = \"admin@streampipes.apache.org\" os.environ[\"SP_API_KEY\"] = \"XXX\" In\u00a0[3]: Copied!
config = StreamPipesClientConfig(\n credential_provider=StreamPipesApiKeyCredentials(),\n host_address=\"localhost\",\n https_disabled=True,\n port=80\n)\nconfig = StreamPipesClientConfig( credential_provider=StreamPipesApiKeyCredentials(), host_address=\"localhost\", https_disabled=True, port=80 ) In\u00a0[4]: Copied!
client = StreamPipesClient(client_config=config)\nclient = StreamPipesClient(client_config=config)
2023-02-24 17:34:25,860 - streampipes.client.client - [INFO] - [client.py:128] [_set_up_logging] - Logging successfully initialized with logging level INFO.\n
As a first step, we want to get an overview about all data available in the data lake. The data is stored as so-called measures, which refer to a data stream stored in the data lake. For his purpose we use the all() method of the dataLakeMeasure endpoint.
data_lake_measures = client.dataLakeMeasureApi.all()\ndata_lake_measures = client.dataLakeMeasureApi.all()
2023-02-24 17:34:25,929 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:167] [_make_request] - Successfully retrieved all resources.\n
So let's see how many measures are available:
In\u00a0[6]: Copied!len(data_lake_measures)\nlen(data_lake_measures) Out[6]:
2
All resources of the StreamPipes Python client support the standard Python expressions. If not, please let us know.
In\u00a0[7]: Copied!data_lake_measures[-1]\ndata_lake_measures[-1] Out[7]:
DataLakeMeasure(element_id='3cb6b5e6f107452483d1fd2ccf4bf9f9', measure_name='test', timestamp_field='s0::timestamp', event_schema=EventSchema(event_properties=[EventProperty(class_name='org.apache.streampipes.model.schema.EventPropertyPrimitive', element_id='sp:eventproperty:EiFnkL', label='Density', description='Denotes the current density of the fluid', runtime_name='density', required=False, domain_properties=['http://schema.org/Number'], property_scope='MEASUREMENT_PROPERTY', index=5, runtime_id=None, runtime_type='http://www.w3.org/2001/XMLSchema#float', measurement_unit=None, value_specification=None), EventProperty(class_name='org.apache.streampipes.model.schema.EventPropertyPrimitive', element_id='sp:eventproperty:ghSkQI', label='Mass Flow', description='Denotes the current mass flow in the sensor', runtime_name='mass_flow', required=False, domain_properties=['http://schema.org/Number'], property_scope='MEASUREMENT_PROPERTY', index=2, runtime_id=None, runtime_type='http://www.w3.org/2001/XMLSchema#float', measurement_unit=None, value_specification=None), EventProperty(class_name='org.apache.streampipes.model.schema.EventPropertyPrimitive', element_id='sp:eventproperty:cQAUry', label='Sensor ID', description='The ID of the sensor', runtime_name='sensorId', required=False, domain_properties=['https://streampipes.org/vocabulary/examples/watertank/v1/hasSensorId'], property_scope='DIMENSION_PROPERTY', index=1, runtime_id=None, runtime_type='http://www.w3.org/2001/XMLSchema#string', measurement_unit=None, value_specification=None), EventProperty(class_name='org.apache.streampipes.model.schema.EventPropertyPrimitive', element_id='sp:eventproperty:pbPMyL', label='Sensor Fault Flags', description='Any fault flags of the sensors', runtime_name='sensor_fault_flags', required=False, domain_properties=['http://schema.org/Boolean'], property_scope='MEASUREMENT_PROPERTY', index=6, runtime_id=None, runtime_type='http://www.w3.org/2001/XMLSchema#boolean', measurement_unit=None, value_specification=None), EventProperty(class_name='org.apache.streampipes.model.schema.EventPropertyPrimitive', element_id='sp:eventproperty:Qmayhw', label='Temperature', description='Denotes the current temperature in degrees celsius', runtime_name='temperature', required=False, domain_properties=['http://schema.org/Number'], property_scope='MEASUREMENT_PROPERTY', index=4, runtime_id=None, runtime_type='http://www.w3.org/2001/XMLSchema#float', measurement_unit='http://qudt.org/vocab/unit#DegreeCelsius', value_specification=ValueSpecification(class_name='org.apache.streampipes.model.schema.QuantitativeValue', element_id=None, min_value=0, max_value=100, step=0.1)), EventProperty(class_name='org.apache.streampipes.model.schema.EventPropertyPrimitive', element_id='sp:eventproperty:YQYhjd', label='Volume Flow', description='Denotes the current volume flow', runtime_name='volume_flow', required=False, domain_properties=['http://schema.org/Number'], property_scope='MEASUREMENT_PROPERTY', index=3, runtime_id=None, runtime_type='http://www.w3.org/2001/XMLSchema#float', measurement_unit=None, value_specification=None)]), pipeline_id=None, pipeline_name=None, pipeline_is_running=False, schema_version='1.1')
To get a more comprehensive overview, you can take a look at the pandas representation:
display(data_lake_measures.to_pandas())\ndisplay(data_lake_measures.to_pandas()) measure_name timestamp_field pipeline_id pipeline_name pipeline_is_running num_event_properties 0 flow-rate s0::timestamp None None False 6 1 test s0::timestamp None None False 6
So far, we have only retrieved metadata about the available data lake measure. In the following, we will access the actual data of the measure flow-rate.
For this purpose, we will use the get() method of the dataLakeMeasure endpoint.
flow_rate_measure = client.dataLakeMeasureApi.get(identifier=\"flow-rate\")\nflow_rate_measure = client.dataLakeMeasureApi.get(identifier=\"flow-rate\")
2023-02-24 17:34:26,020 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:167] [_make_request] - Successfully retrieved all resources.\n
For further processing, the easiest way is to turn the data measure into a pandas DataFrame.
flow_rate_pd = flow_rate_measure.to_pandas()\nflow_rate_pd = flow_rate_measure.to_pandas()
Let's see how many data points we got...
In\u00a0[11]: Copied!len(flow_rate_pd)\nlen(flow_rate_pd) Out[11]:
1000
... and get a first overview
In\u00a0[12]: Copied!flow_rate_pd.describe()\nflow_rate_pd.describe() Out[12]: density mass_flow temperature volume_flow count 1000.000000 1000.000000 1000.000000 1000.000000 mean 45.560337 5.457014 45.480231 5.659558 std 3.201544 3.184959 3.132878 3.122437 min 40.007698 0.004867 40.000992 0.039422 25% 42.819497 2.654101 42.754623 3.021625 50% 45.679264 5.382355 45.435944 5.572553 75% 48.206881 8.183144 48.248473 8.338209 max 50.998310 10.986015 50.964909 10.998676
As a final step, we want to create a plot of both attributes.
In\u00a0[13]: Copied!import matplotlib.pyplot as plt\nflow_rate_pd.plot(y=[\"mass_flow\", \"temperature\"])\nplt.show()\nimport matplotlib.pyplot as plt flow_rate_pd.plot(y=[\"mass_flow\", \"temperature\"]) plt.show()
For data lake measurements, the get() method is even more powerful than simply returning all the data for a given data lake measurement. We will look at a selection of these below. The full list of supported parameters can be found in the docs. Let's start by referring to the graph we created above, where we use only two columns of our data lake measurement. If we already know this, we can directly restrict the queried data to a subset of columns by using the columns parameter. columns takes a list of column names as a comma-separated string:
flow_rate_pd = client.dataLakeMeasureApi.get(identifier=\"flow-rate\", columns=\"mass_flow,temperature\").to_pandas()\nflow_rate_pd\nflow_rate_pd = client.dataLakeMeasureApi.get(identifier=\"flow-rate\", columns=\"mass_flow,temperature\").to_pandas() flow_rate_pd
2023-02-24 17:34:26,492 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:167] [_make_request] - Successfully retrieved all resources.\nOut[14]: timestamp mass_flow temperature 0 2023-02-24T16:19:41.472Z 3.309556 44.448483 1 2023-02-24T16:19:41.482Z 5.608580 40.322033 2 2023-02-24T16:19:41.493Z 7.692881 49.239639 3 2023-02-24T16:19:41.503Z 3.632898 49.933754 4 2023-02-24T16:19:41.513Z 0.711260 50.106617 ... ... ... ... 995 2023-02-24T16:19:52.927Z 1.740114 46.558231 996 2023-02-24T16:19:52.94Z 7.211723 48.048622 997 2023-02-24T16:19:52.952Z 7.770180 48.188026 998 2023-02-24T16:19:52.965Z 4.458602 48.280899 999 2023-02-24T16:19:52.977Z 2.592060 47.505951
1000 rows \u00d7 3 columns
By default, the client returns only the first one thousand records of a Data Lake measurement. This can be changed by passing a concrete value for the limit parameter:
flow_rate_pd = client.dataLakeMeasureApi.get(identifier=\"flow-rate\", limit=10000).to_pandas()\nlen(flow_rate_pd)\nflow_rate_pd = client.dataLakeMeasureApi.get(identifier=\"flow-rate\", limit=10000).to_pandas() len(flow_rate_pd)
2023-02-24 17:34:26,736 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:167] [_make_request] - Successfully retrieved all resources.\nOut[15]:
9528
If you want your data to be selected by time of occurrence rather than quantity, you can specify your time window by passing the start_date and end_date parameters:
from datetime import datetime\nflow_rate_pd = client.dataLakeMeasureApi.get(\n identifier=\"flow-rate\",\n start_date=datetime(year=2023, month=2, day=24, hour=17, minute=21, second=0),\n end_date=datetime(year=2023, month=2, day=24, hour=17, minute=21, second=1),\n ).to_pandas()\nflow_rate_pd.plot(y=[\"mass_flow\", \"temperature\"])\nplt.show()\nfrom datetime import datetime flow_rate_pd = client.dataLakeMeasureApi.get( identifier=\"flow-rate\", start_date=datetime(year=2023, month=2, day=24, hour=17, minute=21, second=0), end_date=datetime(year=2023, month=2, day=24, hour=17, minute=21, second=1), ).to_pandas() flow_rate_pd.plot(y=[\"mass_flow\", \"temperature\"]) plt.show()
2023-02-24 17:34:26,899 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:167] [_make_request] - Successfully retrieved all resources.\n
... from this point on, we leave all future processing of the data up to your creativity. Keep in mind: the general syntax used in this tutorial (all(), to_pandas(), get()) applies to all endpoints and associated resources of the StreamPipes Python client.
If you get further and create exciting stuff with data extracted from StreamPipes please let us know. We are thrilled to see what you as a community will build with the provided client. Furthermore, don't hesitate to discuss feature requests to extend the current functionality with us.
For now, that's all about the StreamPipes client. Read the next tutorial (Getting live data from the StreamPipes data stream) if you are interested in making use of the powerful StreamPipes functions to interact with StreamPipes in an event-based manner.
How do you like this tutorial? We hope you like it and would love to receive some feedback from you. Just go to our GitHub discussion page and let us know your impression. We'll read and react to them all, we promise!
"},{"location":"tutorials/2-extracting-data-from-the-streampipes-data-lake/#Extracting-Data-from-the-StreamPipes-data-lake","title":"Extracting Data from the StreamPipes data lake\u00b6","text":"In the first tutorial (Introduction to the StreamPipes Python client) we took the first steps with the StreamPipes Python client and learned how to set everything up. Now we are ready to get started and want to retrieve some data out of StreamPipes. In this tutorial, we'll focus on the StreamPipes Data Lake, the component where StreamPipes stores data internally. To get started, we'll use the client instance created in the first tutorial.
Note As of now we mainly developed the support for StreamPipes functions using NATS as messaging protocol. Consequently, this setup is tested most and should work flawlessly. Visit our first-steps page to see how to start StreamPipes accordingly. Anyhow, you can also use the other brokers that are currently supported in StreamPipes Python. In case you observe any problems, please reach out to us and file us an issue on GitHub.
In\u00a0[1]: Copied!from streampipes.client import StreamPipesClient\nfrom streampipes.client.config import StreamPipesClientConfig\nfrom streampipes.client.credential_provider import StreamPipesApiKeyCredentials\nfrom streampipes.client import StreamPipesClient from streampipes.client.config import StreamPipesClientConfig from streampipes.client.credential_provider import StreamPipesApiKeyCredentials In\u00a0[\u00a0]: Copied!
# You can install all required libraries for this tutorial with the following command\n%pip install matplotlib ipython streampipes\n# You can install all required libraries for this tutorial with the following command %pip install matplotlib ipython streampipes In\u00a0[2]: Copied!
import os\n\nos.environ[\"SP_USERNAME\"] = \"admin@streampipes.apache.org\"\nos.environ[\"SP_API_KEY\"] = \"XXX\"\n\n# Use this if you work locally:\nos.environ[\"BROKER-HOST\"] = \"localhost\" \nos.environ[\"KAFKA-PORT\"] = \"9094\" # When using Kafka as message broker. If Kafka is not running on localhost, KAFKA_ADVERTISED_LISTENERS should be adjusted to the external address\nimport os os.environ[\"SP_USERNAME\"] = \"admin@streampipes.apache.org\" os.environ[\"SP_API_KEY\"] = \"XXX\" # Use this if you work locally: os.environ[\"BROKER-HOST\"] = \"localhost\" os.environ[\"KAFKA-PORT\"] = \"9094\" # When using Kafka as message broker. If Kafka is not running on localhost, KAFKA_ADVERTISED_LISTENERS should be adjusted to the external address In\u00a0[3]: Copied!
client_config = StreamPipesClientConfig(\n credential_provider=StreamPipesApiKeyCredentials(),\n host_address=\"localhost\",\n port=80,\n https_disabled=True,\n)\nclient = StreamPipesClient(client_config=client_config)\nclient_config = StreamPipesClientConfig( credential_provider=StreamPipesApiKeyCredentials(), host_address=\"localhost\", port=80, https_disabled=True, ) client = StreamPipesClient(client_config=client_config)
2022-12-14 10:43:37,664 - streampipes.client.client - [INFO] - [client.py:127] [_set_up_logging] - Logging successfully initialized with logging level INFO.\n
Now we can have a look at the available data streams. We can choose one or more stream to receive the data from and copy their element_id.
client.dataStreamApi.all().to_pandas()\nclient.dataStreamApi.all().to_pandas()
2022-12-14 10:43:39,944 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:153] [_make_request] - Successfully retrieved all resources.\nOut[4]: element_id name description icon_url app_id includes_assets includes_locales internally_managed measurement_object index ... uri dom num_transport_protocols num_measurement_capability num_application_links num_included_assets num_connected_to num_category num_event_properties num_included_locales 0 urn:streampipes.apache.org:eventstream:HHoidJ Test2 None None False False True None 0 ... urn:streampipes.apache.org:eventstream:HHoidJ None 1 0 0 0 0 0 7 0 1 urn:streampipes.apache.org:eventstream:uPDKLI Test None None False False True None 0 ... urn:streampipes.apache.org:eventstream:uPDKLI None 1 0 0 0 0 0 7 0
2 rows \u00d7 21 columns
Next we can create a StreamPipesFunction. For this we need to implement the 3 following methods:
onServiceStarted is called when the function gets started. There you can use the given meta information of the FunctionContext to initialize the function.onEvent is called when ever a new event arrives. The event contains the live data and you can use the streamId to identify a stream if the function is connected to multiple data streams.onServiceStopped is called when the function gets stopped.For this tutorial we just create a function that saves every new event in a pandas DataFrame and plots the first column of the DataFrame when the function gets stopped.
(If you want to use the same structure as in Java you can overwrite the getFunctionId and requiredStreamIds methods instead of using the FunctionDefinition)
from typing import Dict, Any\nimport pandas as pd\nfrom datetime import datetime\nimport matplotlib.pyplot as plt\nfrom streampipes.functions.function_handler import FunctionHandler\nfrom streampipes.functions.registration import Registration\nfrom streampipes.functions.streampipes_function import StreamPipesFunction\nfrom streampipes.functions.utils.function_context import FunctionContext\nfrom streampipes.model.resource.function_definition import FunctionDefinition, FunctionId\n\nclass ExampleFunction(StreamPipesFunction):\n def __init__(self, function_definition: FunctionDefinition) -> None:\n super().__init__(function_definition)\n # Create the Dataframe to save the live data\n self.df = pd.DataFrame()\n\n def onServiceStarted(self, context: FunctionContext):\n # Get the name of the timestamp field\n for event_property in context.schema[context.streams[0]].event_schema.event_properties:\n if event_property.property_scope == \"HEADER_PROPERTY\":\n self.timestamp = event_property.runtime_name\n\n def onEvent(self, event: Dict[str, Any], streamId: str):\n # Convert the unix timestamp to datetime\n event[self.timestamp] = datetime.fromtimestamp(event[self.timestamp] / 1000)\n # Add every value of the event to the DataFrame\n self.df = pd.concat(\n [self.df, pd.DataFrame({key: [event[key]] for key in event.keys()}).set_index(self.timestamp)]\n )\n\n def onServiceStopped(self):\n # Plot the first column of the Dataframe\n plt.figure(figsize=(10, 5))\n plt.xlabel(self.timestamp)\n plt.ylabel(self.df.columns[0])\n plt.plot(self.df.iloc[:, 0])\n plt.show()\n from typing import Dict, Any import pandas as pd from datetime import datetime import matplotlib.pyplot as plt from streampipes.functions.function_handler import FunctionHandler from streampipes.functions.registration import Registration from streampipes.functions.streampipes_function import StreamPipesFunction from streampipes.functions.utils.function_context import FunctionContext from streampipes.model.resource.function_definition import FunctionDefinition, FunctionId class ExampleFunction(StreamPipesFunction): def __init__(self, function_definition: FunctionDefinition) -> None: super().__init__(function_definition) # Create the Dataframe to save the live data self.df = pd.DataFrame() def onServiceStarted(self, context: FunctionContext): # Get the name of the timestamp field for event_property in context.schema[context.streams[0]].event_schema.event_properties: if event_property.property_scope == \"HEADER_PROPERTY\": self.timestamp = event_property.runtime_name def onEvent(self, event: Dict[str, Any], streamId: str): # Convert the unix timestamp to datetime event[self.timestamp] = datetime.fromtimestamp(event[self.timestamp] / 1000) # Add every value of the event to the DataFrame self.df = pd.concat( [self.df, pd.DataFrame({key: [event[key]] for key in event.keys()}).set_index(self.timestamp)] ) def onServiceStopped(self): # Plot the first column of the Dataframe plt.figure(figsize=(10, 5)) plt.xlabel(self.timestamp) plt.ylabel(self.df.columns[0]) plt.plot(self.df.iloc[:, 0]) plt.show() Now we can start the function. First we create an instance of the ExampleFunction and insert the element_id of the stream which data we want to consume. Then we have to register this function and we can start all functions by initializing the FunctionHandler. (it's also possible to register multiple functions with .register(...).register(...))
example_function = ExampleFunction(\n FunctionDefinition(\n function_id=FunctionId(id=\"example-function\"),\n consumed_streams=[\"urn:streampipes.apache.org:eventstream:uPDKLI\"]\n )\n)\n\nregistration = Registration()\nregistration.register(example_function)\n\nfunction_handler = FunctionHandler(registration, client)\nfunction_handler.initializeFunctions()\nexample_function = ExampleFunction( FunctionDefinition( function_id=FunctionId(id=\"example-function\"), consumed_streams=[\"urn:streampipes.apache.org:eventstream:uPDKLI\"] ) ) registration = Registration() registration.register(example_function) function_handler = FunctionHandler(registration, client) function_handler.initializeFunctions()
2022-12-14 10:43:42,810 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:153] [_make_request] - Successfully retrieved all resources.\n2022-12-14 10:43:42,812 - streampipes.functions.function_handler - [INFO] - [function_handler.py:82] [initializeFunctions] - Using NatsBroker for <__main__.ExampleFunction object at 0x000001F2EF298D90>\n
The DataFrame of the ExampleFunction gets updated whenever a new event arrives. Let's test this by executing the cell below.
import asyncio\nfrom asyncio.exceptions import CancelledError\nfrom IPython.display import clear_output\n\nwhile True:\n clear_output(wait=True)\n display(example_function.df)\n try:\n await asyncio.sleep(1)\n except CancelledError:\n break\nimport asyncio from asyncio.exceptions import CancelledError from IPython.display import clear_output while True: clear_output(wait=True) display(example_function.df) try: await asyncio.sleep(1) except CancelledError: break mass_flow density volume_flow sensor_fault_flags temperature sensorId timestamp 2022-12-14 10:43:43.357 10.955496 47.546290 1.001985 False 44.993413 flowrate02 2022-12-14 10:43:44.371 6.499040 44.392069 2.034402 False 41.232352 flowrate02 2022-12-14 10:43:45.382 10.168300 41.192146 9.724287 False 46.812779 flowrate02 2022-12-14 10:43:46.395 10.849059 50.086308 5.832691 False 45.860412 flowrate02 2022-12-14 10:43:47.410 3.081855 47.254246 8.860531 False 50.505801 flowrate02 ... ... ... ... ... ... ... 2022-12-14 10:44:43.920 1.803572 41.978894 10.294002 False 47.820239 flowrate02 2022-12-14 10:44:44.932 1.967062 42.212883 3.237440 False 49.047258 flowrate02 2022-12-14 10:44:45.934 4.457819 47.561256 0.315024 False 40.223413 flowrate02 2022-12-14 10:44:46.949 8.745343 46.346891 7.439090 False 41.982529 flowrate02 2022-12-14 10:44:47.950 5.828744 47.679720 6.307405 False 42.100354 flowrate02
65 rows \u00d7 6 columns
The while loop just displays the the DataFrame every second until the cell is stopped. We could achieve the same result manually by executing example_function.df repeatedly.
You can stop the functions whenever you want by executing the command below.
In\u00a0[\u00a0]: Copied!function_handler.disconnect()\nfunction_handler.disconnect()
2022-12-14 10:44:53,309 - streampipes.functions.broker.nats_broker - [INFO] - [nats_broker.py:67] [disconnect] - Stopped connection to stream: urn:streampipes.apache.org:eventstream:uPDKLI\n
That's enough for this tutorial. Now you can try to write your own StreamPipesFunction. All you need to do is creating a new class, implementing the 4 required methods and registering the function.
Want to see more exciting use cases you can achieve with StreamPipes functions in Python? Then don't hesitate and jump to our next tutorial on applying online machine learning algorithms to StreamPipes data streams with River.
How do you like this tutorial? We hope you like it and would love to receive some feedback from you. Just go to our GitHub discussion page and let us know your impression. We'll read and react to them all, we promise!
"},{"location":"tutorials/3-getting-live-data-from-the-streampipes-data-stream/#Getting-live-data-from-the-StreamPipes-data-stream","title":"Getting live data from the StreamPipes data stream\u00b6","text":"In the last tutorial (Extracting Data from the StreamPipes data lake) we learned how to extract the stored data from a StreamPipes data lake. This tutorial is about the StreamPipes data stream and shows how to get the live data from StreamPipes into Python. Therefore, we first create the client instance as before.
from streampipes.client import StreamPipesClient\nfrom streampipes.client.config import StreamPipesClientConfig\nfrom streampipes.client.credential_provider import StreamPipesApiKeyCredentials\nfrom streampipes.client import StreamPipesClient from streampipes.client.config import StreamPipesClientConfig from streampipes.client.credential_provider import StreamPipesApiKeyCredentials In\u00a0[\u00a0]: Copied!
# you can install all required dependencies for this tutorial by executing the following command\n%pip install river streampipes\n# you can install all required dependencies for this tutorial by executing the following command %pip install river streampipes In\u00a0[2]: Copied!
import os\n\nos.environ[\"SP_USERNAME\"] = \"admin@streampipes.apache.org\"\nos.environ[\"SP_API_KEY\"] = \"XXX\"\n\n# Use this if you work locally:\nos.environ[\"BROKER-HOST\"] = \"localhost\" \nos.environ[\"KAFKA-PORT\"] = \"9094\" # When using Kafka as message broker\nimport os os.environ[\"SP_USERNAME\"] = \"admin@streampipes.apache.org\" os.environ[\"SP_API_KEY\"] = \"XXX\" # Use this if you work locally: os.environ[\"BROKER-HOST\"] = \"localhost\" os.environ[\"KAFKA-PORT\"] = \"9094\" # When using Kafka as message broker In\u00a0[3]: Copied!
client_config = StreamPipesClientConfig(\n credential_provider=StreamPipesApiKeyCredentials(),\n host_address=\"localhost\",\n port=80,\n https_disabled=True,\n)\nclient = StreamPipesClient(client_config=client_config)\nclient_config = StreamPipesClientConfig( credential_provider=StreamPipesApiKeyCredentials(), host_address=\"localhost\", port=80, https_disabled=True, ) client = StreamPipesClient(client_config=client_config)
2023-01-27 16:04:24,784 - streampipes.client.client - [INFO] - [client.py:128] [_set_up_logging] - Logging successfully initialized with logging level INFO.\nIn\u00a0[4]: Copied!
client.dataStreamApi.all().to_pandas()\nclient.dataStreamApi.all().to_pandas()
2023-01-27 16:04:28,212 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:163] [_make_request] - Successfully retrieved all resources.\nOut[4]: element_id name description icon_url app_id includes_assets includes_locales internally_managed measurement_object index ... dom rev num_transport_protocols num_measurement_capability num_application_links num_included_assets num_connected_to num_category num_event_properties num_included_locales 0 sp:spdatastream:xboBFK Test None None False False True None 0 ... None 5-558c861debc745e1ebae29a266a8bdb9 1 0 0 0 0 0 7 0 1 urn:streampipes.apache.org:eventstream:Wgyrse Test File None None False False True None 0 ... None 4-66548b6b84287011b7cec0876ef82baf 1 0 0 0 0 0 2 0
2 rows \u00d7 22 columns
In\u00a0[5]: Copied!from river import cluster, compose, preprocessing\nfrom streampipes.function_zoo.river_function import OnlineML\nfrom streampipes.functions.utils.data_stream_generator import RuntimeType\n\nk_means = compose.Pipeline(\n (\"drop_features\", compose.Discard(\"sensorId\", \"timestamp\")),\n (\"scale\", preprocessing.StandardScaler()),\n (\"k_means\", cluster.KMeans(n_clusters=2)),\n)\n\nclustering = OnlineML(\n client=client, stream_ids=[\"sp:spdatastream:xboBFK\"], model=k_means, prediction_type=RuntimeType.INTEGER.value\n)\nclustering.start()\nfrom river import cluster, compose, preprocessing from streampipes.function_zoo.river_function import OnlineML from streampipes.functions.utils.data_stream_generator import RuntimeType k_means = compose.Pipeline( (\"drop_features\", compose.Discard(\"sensorId\", \"timestamp\")), (\"scale\", preprocessing.StandardScaler()), (\"k_means\", cluster.KMeans(n_clusters=2)), ) clustering = OnlineML( client=client, stream_ids=[\"sp:spdatastream:xboBFK\"], model=k_means, prediction_type=RuntimeType.INTEGER.value ) clustering.start()
2023-01-27 16:04:35,599 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:163] [_make_request] - Successfully retrieved all resources.\n2023-01-27 16:04:35,599 - streampipes.functions.function_handler - [INFO] - [function_handler.py:64] [initializeFunctions] - Create output data stream \"sp:spdatastream:cwKPoo\" for the function \"65cf8b86-bcdf-433e-a1c7-3e920eab55d0\"\n2023-01-27 16:04:37,766 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:163] [_make_request] - Successfully retrieved all resources.\n2023-01-27 16:04:37,767 - streampipes.functions.function_handler - [INFO] - [function_handler.py:78] [initializeFunctions] - Using NatsBroker for RiverFunction\n
2023-01-27 16:04:37,791 - streampipes.functions.broker.nats_broker - [INFO] - [nats_broker.py:48] [_makeConnection] - Connected to NATS at localhost:4222\n2023-01-27 16:04:37,791 - streampipes.functions.broker.nats_broker - [INFO] - [nats_broker.py:48] [_makeConnection] - Connected to NATS at localhost:4222\n2023-01-27 16:04:37,792 - streampipes.functions.broker.nats_broker - [INFO] - [nats_broker.py:58] [createSubscription] - Subscribed to stream: sp:spdatastream:xboBFK\nIn\u00a0[6]: Copied!
clustering.set_learning(False)\nclustering.set_learning(False) In\u00a0[\u00a0]: Copied!
clustering.stop()\nclustering.stop()
2023-01-27 16:04:57,303 - streampipes.functions.broker.nats_broker - [INFO] - [nats_broker.py:82] [disconnect] - Stopped connection to stream: sp:spdatastream:xboBFK\n2023-01-27 16:04:57,304 - streampipes.functions.broker.nats_broker - [INFO] - [nats_broker.py:82] [disconnect] - Stopped connection to stream: sp:spdatastream:cwKPoo\nIn\u00a0[\u00a0]: Copied!
import pickle\nfrom river import compose, tree\nfrom streampipes.function_zoo.river_function import OnlineML\nfrom streampipes.functions.utils.data_stream_generator import RuntimeType\n\nhoeffding_tree = compose.Pipeline(\n (\"drop_features\", compose.Discard(\"sensorId\", \"timestamp\")),\n (\"hoeffding_tree\", tree.HoeffdingTreeRegressor(grace_period=5)),\n)\n\n\ndef draw_tree(self, event, streamId):\n\"\"\"Draw the tree and save the image.\"\"\"\n if self.learning:\n if self.model[1].n_nodes != None:\n self.model[1].draw().render(\"hoeffding_tree\", format=\"png\", cleanup=True)\n\n\ndef save_model(self):\n\"\"\"Save the trained model.\"\"\"\n with open(\"hoeffding_tree.pkl\", \"wb\") as f:\n pickle.dump(self.model, f)\n\n\nregressor = OnlineML(\n client=client,\n stream_ids=[\"sp:spdatastream:xboBFK\"],\n model=hoeffding_tree,\n prediction_type=RuntimeType.FLOAT.value,\n supervised=True,\n target_label=\"temperature\",\n on_event=draw_tree,\n on_stop=save_model,\n)\nregressor.start()\nimport pickle from river import compose, tree from streampipes.function_zoo.river_function import OnlineML from streampipes.functions.utils.data_stream_generator import RuntimeType hoeffding_tree = compose.Pipeline( (\"drop_features\", compose.Discard(\"sensorId\", \"timestamp\")), (\"hoeffding_tree\", tree.HoeffdingTreeRegressor(grace_period=5)), ) def draw_tree(self, event, streamId): \"\"\"Draw the tree and save the image.\"\"\" if self.learning: if self.model[1].n_nodes != None: self.model[1].draw().render(\"hoeffding_tree\", format=\"png\", cleanup=True) def save_model(self): \"\"\"Save the trained model.\"\"\" with open(\"hoeffding_tree.pkl\", \"wb\") as f: pickle.dump(self.model, f) regressor = OnlineML( client=client, stream_ids=[\"sp:spdatastream:xboBFK\"], model=hoeffding_tree, prediction_type=RuntimeType.FLOAT.value, supervised=True, target_label=\"temperature\", on_event=draw_tree, on_stop=save_model, ) regressor.start() In\u00a0[9]: Copied!
regressor.set_learning(False)\nregressor.set_learning(False) In\u00a0[\u00a0]: Copied!
regressor.stop()\nregressor.stop() In\u00a0[\u00a0]: Copied!
import pickle\nfrom river import compose, tree\nfrom streampipes.function_zoo.river_function import OnlineML\nfrom streampipes.functions.utils.data_stream_generator import RuntimeType\n\ndecision_tree = compose.Pipeline(\n (\"drop_features\", compose.Discard(\"sensorId\", \"timestamp\")),\n (\"decision_tree\", tree.ExtremelyFastDecisionTreeClassifier(grace_period=5)),\n)\n\n\ndef draw_tree(self, event, streamId):\n\"\"\"Draw the tree and save the image.\"\"\"\n if self.learning:\n if self.model[1].n_nodes != None:\n self.model[1].draw().render(\"decicion_tree\", format=\"png\", cleanup=True)\n\n\ndef save_model(self):\n\"\"\"Save the trained model.\"\"\"\n with open(\"decision_tree.pkl\", \"wb\") as f:\n pickle.dump(self.model, f)\n\n\nclassifier = OnlineML(\n client=client,\n stream_ids=[\"sp:spdatastream:xboBFK\"],\n model=decision_tree,\n prediction_type=RuntimeType.BOOLEAN.value,\n supervised=True,\n target_label=\"sensor_fault_flags\",\n on_event=draw_tree,\n on_stop=save_model,\n)\nclassifier.start()\nimport pickle from river import compose, tree from streampipes.function_zoo.river_function import OnlineML from streampipes.functions.utils.data_stream_generator import RuntimeType decision_tree = compose.Pipeline( (\"drop_features\", compose.Discard(\"sensorId\", \"timestamp\")), (\"decision_tree\", tree.ExtremelyFastDecisionTreeClassifier(grace_period=5)), ) def draw_tree(self, event, streamId): \"\"\"Draw the tree and save the image.\"\"\" if self.learning: if self.model[1].n_nodes != None: self.model[1].draw().render(\"decicion_tree\", format=\"png\", cleanup=True) def save_model(self): \"\"\"Save the trained model.\"\"\" with open(\"decision_tree.pkl\", \"wb\") as f: pickle.dump(self.model, f) classifier = OnlineML( client=client, stream_ids=[\"sp:spdatastream:xboBFK\"], model=decision_tree, prediction_type=RuntimeType.BOOLEAN.value, supervised=True, target_label=\"sensor_fault_flags\", on_event=draw_tree, on_stop=save_model, ) classifier.start() In\u00a0[12]: Copied!
classifier.set_learning(False)\nclassifier.set_learning(False) In\u00a0[\u00a0]: Copied!
classifier.stop()\nclassifier.stop()
That's already it! Isn't it truly easy to apply Online ML with StreamPipes and River? Please go ahead and apply it to your own use cases. We would be happy to hear about them!
Want to see more exciting use cases you can achieve with StreamPipes functions in Python? Then don\u2019t hesitate and jump to our next tutorial on using interoperable machine learning algorithm models with StreamPipes Python and ONNX.
How do you like this tutorial? We hope you like it and would love to receive some feedback from you. Just go to our GitHub discussion page and let us know your impression. We'll read and react to them all, we promise!
"},{"location":"tutorials/4-using-online-machine-learning-on-a-streampipes-data-stream/#Using-Online-Machine-Learning-on-a-StreamPipes-data-stream","title":"Using Online Machine Learning on a StreamPipes data stream\u00b6","text":"The last tutorial (Getting live data from the StreamPipes data stream) showed how we can connect to a data stream, and it would be possible to use Online Machine Learning with this approach and train a model with the incoming events at the onEvent method. However, the StreamPipes client also provides an easier way to do this with the use of the River library for Online Machine Learning. We will have a look at this now.
After we configured the client as usual, we can start with the new part. The approach is straight forward and you can start with the ML part in just 3 steps:
OnlineML wrapper to fit to your model and insert the client and required data stream ids.A StreamPipesFunction is then started, which trains the model for each new event. It also creates an output data stream which will send the prediction of the model back to StreamPipes. This output stream can be seen when creating a new pipeline and can be used like every other data source. So you can use it in a pipeline and save the predictions in a Data Lake. You can also stop and start the training with the method set_learning. To stop the whole function use the stop methode and if you want to delete the output stream entirely, you can go to the Pipeline Element Installer in StreamPipes and uninstall it.
Now let's take a look at some examples. If you want to execute the examples below you have to create an adapter for the Machine Data Simulator, select the flowrate sensor and insert the stream id of this stream.
%pip install git+https://github.com/apache/streampipes.git#subdirectory=streampipes-client-python\n%pip install scikit-learn==1.4.0 skl2onnx==1.16.0 onnxruntime==1.17.1 matplotlib==3.8.3\n%pip install git+https://github.com/apache/streampipes.git#subdirectory=streampipes-client-python %pip install scikit-learn==1.4.0 skl2onnx==1.16.0 onnxruntime==1.17.1 matplotlib==3.8.3 In\u00a0[8]: Copied!
import os\nfrom streampipes.client import StreamPipesClient\nfrom streampipes.client.config import StreamPipesClientConfig\nfrom streampipes.client.credential_provider import StreamPipesApiKeyCredentials\n\nos.environ[\"BROKER-HOST\"] = \"localhost\"\nos.environ[\"KAFKA-PORT\"] = \"9094\" # When using Kafka as message broker\n\nconfig = StreamPipesClientConfig(\n credential_provider=StreamPipesApiKeyCredentials(\n username=\"admin@streampipes.apache.org\",\n api_key=\"TOKEN\",\n ),\n host_address=\"localhost\",\n https_disabled=True,\n port=80\n)\n\nclient = StreamPipesClient(client_config=config)\nimport os from streampipes.client import StreamPipesClient from streampipes.client.config import StreamPipesClientConfig from streampipes.client.credential_provider import StreamPipesApiKeyCredentials os.environ[\"BROKER-HOST\"] = \"localhost\" os.environ[\"KAFKA-PORT\"] = \"9094\" # When using Kafka as message broker config = StreamPipesClientConfig( credential_provider=StreamPipesApiKeyCredentials( username=\"admin@streampipes.apache.org\", api_key=\"TOKEN\", ), host_address=\"localhost\", https_disabled=True, port=80 ) client = StreamPipesClient(client_config=config)
2024-03-26 10:21:38,538 - streampipes.client.client - [INFO] - [client.py:198] [_set_up_logging] - Logging successfully initialized with logging level INFO.\n2024-03-26 10:21:38,632 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:164] [_make_request] - Successfully retrieved all resources.\n2024-03-26 10:21:38,634 - streampipes.client.client - [INFO] - [client.py:171] [_get_server_version] - The StreamPipes version was successfully retrieved from the backend: 0.95.0. By means of that, authentication via the provided credentials is also tested successfully.\n
The main objective of this tutorial is to demonstrate how to make predictions with an existing and pre-trained ML model using a StreamPipes function and ONNX. Therefore, you can skip the following sections on use case and model training if you already have an existing ONNX model and are only interested in applying it using StreamPipes.
In\u00a0[9]: Copied!flowrate_df = client.dataLakeMeasureApi.get(\"flow-rate\").to_pandas()\nX = flowrate_df[\"volume_flow\"].values.reshape(-1, 1).astype(\"float32\")\nflowrate_df = client.dataLakeMeasureApi.get(\"flow-rate\").to_pandas() X = flowrate_df[\"volume_flow\"].values.reshape(-1, 1).astype(\"float32\")
2024-03-26 10:21:48,582 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:164] [_make_request] - Successfully retrieved all resources.\n
Let's fit the model to the data:
In\u00a0[14]: Copied!from sklearn.ensemble import IsolationForest\n\nmodel = IsolationForest(contamination=0.01)\nmodel.fit(X)\nfrom sklearn.ensemble import IsolationForest model = IsolationForest(contamination=0.01) model.fit(X) Out[14]:
IsolationForest(contamination=0.01)In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.\u00a0\u00a0IsolationForest?Documentation for IsolationForestiFitted
IsolationForest(contamination=0.01)
The contamination parameter models the proportion of outliers in the data. See the scikit-learn documentation for more information.
Before we convert the model to an ONNX representation, let's do a quick visual analysis of the model results:
In\u00a0[15]: Copied!import matplotlib.pyplot as plt\n\nflowrate_df[\"anomaly\"] = model.predict(X)\n\nfig, ax = plt.subplots(figsize=(10, 6))\nanomalies = flowrate_df.loc[flowrate_df[\"anomaly\"] == -1, [\"volume_flow\"]]\nax.plot(flowrate_df.index, flowrate_df['volume_flow'], color='black', label='volume_flow')\nax.scatter(anomalies.index, anomalies['volume_flow'], color='red', label='Anomaly')\nplt.legend()\nplt.show()\nimport matplotlib.pyplot as plt flowrate_df[\"anomaly\"] = model.predict(X) fig, ax = plt.subplots(figsize=(10, 6)) anomalies = flowrate_df.loc[flowrate_df[\"anomaly\"] == -1, [\"volume_flow\"]] ax.plot(flowrate_df.index, flowrate_df['volume_flow'], color='black', label='volume_flow') ax.scatter(anomalies.index, anomalies['volume_flow'], color='red', label='Anomaly') plt.legend() plt.show()
Okay, that looks quite reasonable so lets covert the model to an ONNX representation so that we can make use of it later.
In\u00a0[16]: Copied!from onnxconverter_common import FloatTensorType\nfrom skl2onnx import to_onnx\n\nmodel_onnx = to_onnx(\n model,\n initial_types=[('input', FloatTensorType([None, X.shape[1]]))],\n target_opset={'ai.onnx.ml': 3, 'ai.onnx': 15, '': 15}\n)\n\nwith open(\"isolation_forest.onnx\", \"wb\") as f:\n f.write(model_onnx.SerializeToString())\n from onnxconverter_common import FloatTensorType from skl2onnx import to_onnx model_onnx = to_onnx( model, initial_types=[('input', FloatTensorType([None, X.shape[1]]))], target_opset={'ai.onnx.ml': 3, 'ai.onnx': 15, '': 15} ) with open(\"isolation_forest.onnx\", \"wb\") as f: f.write(model_onnx.SerializeToString()) In\u00a0[23]: Copied! import numpy as np\nimport onnxruntime as rt\n\nfrom streampipes.functions.broker.broker_handler import get_broker_description\nfrom streampipes.functions.streampipes_function import StreamPipesFunction\nfrom streampipes.functions.utils.data_stream_generator import create_data_stream, RuntimeType\nfrom streampipes.functions.utils.function_context import FunctionContext\nfrom streampipes.model.resource import FunctionDefinition, DataStream\n\nfrom typing import Dict, Any, List\n\n\nclass ONNXFunction(StreamPipesFunction):\n\n def __init__(self, feature_names: list[str], input_stream: DataStream):\n output_stream = create_data_stream(\n name=\"flowrate-prediction\",\n attributes={\n \"is_anomaly\": RuntimeType.BOOLEAN.value\n },\n broker=get_broker_description(input_stream)\n )\n\n function_definition = FunctionDefinition(\n consumed_streams=[input_stream.element_id]\n ).add_output_data_stream(output_stream)\n\n self.feature_names = feature_names\n self.input_name = None\n self.output_name = None\n self.session = None\n\n super().__init__(function_definition=function_definition)\n\n def onServiceStarted(self, context: FunctionContext) -> None:\n self.session = rt.InferenceSession(\n path_or_bytes=\"isolation_forest.onnx\",\n providers=rt.get_available_providers(),\n )\n self.input_name = self.session.get_inputs()[0].name\n self.output_name = self.session.get_outputs()[0].name\n\n def onEvent(self, event: Dict[str, Any], streamId: str) -> None:\n feature_vector = []\n for feature in self.feature_names:\n feature_vector.append(event[feature])\n\n prediction = self.session.run(\n [self.output_name],\n {self.input_name: np.expand_dims(np.array(feature_vector), axis=0).astype(\"float32\")}\n )[0]\n\n output = {\n \"is_anomaly\": int(prediction[0]) == -1\n }\n\n self.add_output(\n stream_id=self.function_definition.get_output_stream_ids()[0],\n event=output\n )\n\n def onServiceStopped(self) -> None:\n pass\n import numpy as np import onnxruntime as rt from streampipes.functions.broker.broker_handler import get_broker_description from streampipes.functions.streampipes_function import StreamPipesFunction from streampipes.functions.utils.data_stream_generator import create_data_stream, RuntimeType from streampipes.functions.utils.function_context import FunctionContext from streampipes.model.resource import FunctionDefinition, DataStream from typing import Dict, Any, List class ONNXFunction(StreamPipesFunction): def __init__(self, feature_names: list[str], input_stream: DataStream): output_stream = create_data_stream( name=\"flowrate-prediction\", attributes={ \"is_anomaly\": RuntimeType.BOOLEAN.value }, broker=get_broker_description(input_stream) ) function_definition = FunctionDefinition( consumed_streams=[input_stream.element_id] ).add_output_data_stream(output_stream) self.feature_names = feature_names self.input_name = None self.output_name = None self.session = None super().__init__(function_definition=function_definition) def onServiceStarted(self, context: FunctionContext) -> None: self.session = rt.InferenceSession( path_or_bytes=\"isolation_forest.onnx\", providers=rt.get_available_providers(), ) self.input_name = self.session.get_inputs()[0].name self.output_name = self.session.get_outputs()[0].name def onEvent(self, event: Dict[str, Any], streamId: str) -> None: feature_vector = [] for feature in self.feature_names: feature_vector.append(event[feature]) prediction = self.session.run( [self.output_name], {self.input_name: np.expand_dims(np.array(feature_vector), axis=0).astype(\"float32\")} )[0] output = { \"is_anomaly\": int(prediction[0]) == -1 } self.add_output( stream_id=self.function_definition.get_output_stream_ids()[0], event=output ) def onServiceStopped(self) -> None: pass Let's dive a little deeper into the different parts of the function
__init__: First, we need to take care of the data stream that is required to send the predictions from our function to StreamPipes. Thus, we create a dedicated output data stream which we need to provide with the attributes our event will consist of (a timestamp attribute is always added automatically). This output data stream needs to be registered at the function definition which is to be passed to the parent class. Lastly, we need to define some instance variables that are mainly required for the ONNX runtime.
onServiceStarted: Here we prepare the ONNX runtime session by creating an InferenceSession and retrieving the corresponding configuration parameters.
onEvent: Following the parameter names specified by self.feature_names, we extract all feature values from the current event. Subsequently, the corresponding feature vector is transmitted to the ONNX runtime session. The resulting prediction is then converted into our output event, where a value of -1 signifies an anomaly. Finally, the generated output event is forwarded to StreamPipes.
Having the function code in place, we can start the function with the following:
In\u00a0[25]: Copied!from streampipes.functions.registration import Registration\nfrom streampipes.functions.function_handler import FunctionHandler\n\nstream = [\n stream\n for stream\n in client.dataStreamApi.all()\n if stream.name == \"flow-rate\"\n][0]\n\nfunction = ONNXFunction(\n feature_names=[\"volume_flow\"],\n input_stream=stream\n)\n\nregistration = Registration()\nregistration.register(function)\nfunction_handler = FunctionHandler(registration, client)\nfunction_handler.initializeFunctions()\nfrom streampipes.functions.registration import Registration from streampipes.functions.function_handler import FunctionHandler stream = [ stream for stream in client.dataStreamApi.all() if stream.name == \"flow-rate\" ][0] function = ONNXFunction( feature_names=[\"volume_flow\"], input_stream=stream ) registration = Registration() registration.register(function) function_handler = FunctionHandler(registration, client) function_handler.initializeFunctions()
2024-03-26 12:39:50,443 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:164] [_make_request] - Successfully retrieved all resources.\n2024-03-26 12:39:50,502 - streampipes.functions.function_handler - [INFO] - [function_handler.py:76] [initializeFunctions] - The data stream could not be created.\n2024-03-26 12:39:50,503 - streampipes.functions.function_handler - [INFO] - [function_handler.py:78] [initializeFunctions] - This is due to the fact that this data stream already exists. Continuing with the existing data stream.\n2024-03-26 12:39:50,503 - streampipes.functions.function_handler - [INFO] - [function_handler.py:84] [initializeFunctions] - Using output data stream 'sp:spdatastream:flowrate-prediction' for function '7c06fa31-9534-4f91-9c50-b7a3607ec3dc'\n2024-03-26 12:39:50,548 - streampipes.endpoint.endpoint - [INFO] - [endpoint.py:164] [_make_request] - Successfully retrieved all resources.\n2024-03-26 12:39:50,549 - streampipes.functions.function_handler - [INFO] - [function_handler.py:100] [initializeFunctions] - Using KafkaConsumer for ONNXFunction\n
We can now access the live values of the prediction in the StreamPipes UI, e.g., in the pipeline editor.
That's already it. We hope this tutorial serves as an illustration how ML models can be utilized in StreamPipes with the help of ONNX.
How do you like this tutorial? We hope you like it and would love to receive some feedback from you. Just go to our GitHub discussion page and let us know your impression. We'll read and react to them all, we promise!
"},{"location":"tutorials/5-applying-interoperable-machine-learning-in-streampipes/#Applying-Interoperable-Machine-Learning-in-StreamPipes","title":"Applying Interoperable Machine Learning in StreamPipes\u00b6","text":"The last tutorial (Using Online Machine Learning on a StreamPipes data stream) demonstrated how patterns in streaming data can be learned online. In contrast, this tutorial demonstrates how one can apply a pre-trained machine learning (ML) model to a StreamPipes data stream making use of ONNX. We will show how StreamPipes can be used for both: extracting historical data for training purposes and using model inference on live data with a pre-trained model.
"},{"location":"tutorials/5-applying-interoperable-machine-learning-in-streampipes/#Preparation","title":"Preparation\u00b6","text":"The following lines configure the client and establish a connection to the StreamPipes instance. If you're not familiar with it or anything is unclear, please have a look at our first tutorial.
"},{"location":"tutorials/5-applying-interoperable-machine-learning-in-streampipes/#Machine-Learning-Use-Case","title":"Machine Learning Use Case\u00b6","text":"In this tutorial, we will use data generated by the Machine Data Simulator adapter. More specifically, we will focus on the flowrate data, which consists of various sensor values coming from a water pipe system. Our goal is keep an eye on the parameter volume_flow, which represents the current volume flow in cubic meters/second. For this parameter, we want to detect anomalies that could indicate problems such as leaks, blockages, etc.
To get the concerned data, we simply need to create an instance of the machine data simulator and persist the data in the data lake:
If you choose to perform the model training step yourself, you will need to wait approximately 15 minutes for enough data to be available for model training. If you want to speed this up, you can configure a lower wait time when creating the adapter. Please be aware that this also influences the inference scenario.
"},{"location":"tutorials/5-applying-interoperable-machine-learning-in-streampipes/#Model-Training-with-Historic-Data","title":"Model Training with Historic Data\u00b6","text":"As said above, the aim of our model is to detect anomalies of the volume_flow parameter. For this task, we will use Isolation Forests. Please note that the focus of the tutorial is not on training the model, so please be patient even though the training is very simplified and lacks important preparation steps such as standardization.
As a first step, lets query the flowrate data from the StreamPipes data lake and extract the values of volume_flow as a feature:
Utilizing a pre-trained model within StreamPipes becomes seamless with the ONNX interoperability standard, enabling effortless application of your existing model on live data streams.
Interacting with live data from StreamPipes is facilitated through StreamPipes functions. Below, we'll create a Python StreamPipes function that leverages an ONNX model to generate predictions for each incoming event, making the results accessible as a data stream within StreamPipes for subsequent steps.
So let's create an ONNXFunction that is capable of applying a model in ONNX representation to a StreamPipes data stream. If you'd like to read more details about how functions are defined, refer to our third tutorial.