Skip to content

Latest commit

 

History

History
246 lines (190 loc) · 7.63 KB

README.md

File metadata and controls

246 lines (190 loc) · 7.63 KB
Raw

Set up

  1. Pull the mage-ai repository and cd into it.
  2. Run Mage in a development Docker container using ./scripts/dev.sh [PROJECT NAME]. This starts Mage and allows us to make changes in realtime. See this page for more details.
  3. Open another terminal and run: 
    docker exec -it mage-ai-server-1 bash
    This will open a shell in the Docker container and allow us to interact with the integrations.
  4. Uninstall the existing mage-integrations package using pip: 
    pip3 uninstall -y mage-integrations
  5. cd into mage_integrations/. 
    cd mage_integrations/

Testing sync locally

Run

touch ./mage_integrations/TEST_CATALOG.json &&
touch ./mage_integrations/TEST_CONFIG_S.json &&
touch ./mage_integrations/TEST_CONFIG_D.json &&
touch ./mage_integrations/TEST_STATE.json &&
touch ./mage_integrations/TEST_OUTPUT &&
echo "{}" >> ./mage_integrations/TEST_STATE.json

To create the following files:

TEST_CATALOG.json
TEST_CONFIG_S.json
TEST_CONFIG_D.json
TEST_STATE.json
TEST_OUTPUT

We'll be using these files to test the integration.

Set up source config

Populate TEST_CONFIG_S.json with a sample configuration, this is found at:

mage_integrations/mage_integrations/sources/[INTEGRATION]/templates/config.json

For the GitHub integration, this is:

{
  "access_token": "abcdefghijklmnopqrstuvwxyz1234567890ABCD",
  "repository": "mage-ai/mage-ai",
  "start_date": "2021-01-01T00:00:00Z",
  "request_timeout": 300,
  "base_url": "https://api.github.com"
}

Discover streams

Run the following command to discover streams for your source:

python3 mage_integrations/sources/[INTEGRATION]/__init__.py \
  --config mage_integrations/TEST_CONFIG_S.json \
  --discover \
  --discover_streams

Now, you should see a list of streams that you can sync from the source. Grab a few of interest!

The output from GitHub looks like this:

[
    {
        "stream": "commits",
        "tap_stream_id": "commits"
    },
    {
        "stream": "comments",
        "tap_stream_id": "comments"
    },
    ...
]

Get schema for streams

Now test grabbing schemas for a few streams above and output the data to our catalog file. This should be passed, in string format, as a list of strings, e.g. '["commits", "comments"]'.

python3 mage_integrations/sources/[INTEGRATION]/__init__.py \
  --config mage_integrations/TEST_CONFIG_S.json \
  --discover \
  --selected_streams SCHEMAS > mage_integrations/TEST_CATALOG.json

For example, for the GitHub source:

python3 mage_integrations/sources/github/__init__.py \
  --config mage_integrations/TEST_CONFIG_S.json \
  --discover \
  --selected_streams '["commits"]' > mage_integrations/TEST_CATALOG.json

Your catalog will now contain the schemas for the streams you selected. We need to enable a schema now— this is usually handled by the Mage UI, but we can do it ourselves.

For each stream in TEST_CATALOG.json, find the nested metadata key and add a "selected": true:

...
"stream": "commits",
"metadata": [
  {
    "breadcrumb": [],
    "metadata": {
      "table-key-properties": [
        "sha"
      ],
      "forced-replication-method": "INCREMENTAL",
      "valid-replication-keys": "updated_at",
      "inclusion": "available",
      "selected": true
    }
  },
...

Additonally, also add to add "selected": true to at least one column in each stream for SQL sources.

{
  "stream": "commits",
  "tap_stream_id": "commits",
  "selected": true
}

Test stream execution and save to output file

Finally! It's time to test our stream execution. Run the following command to execute the stream and save the output to a file:

python3 mage_integrations/sources/[INTEGRATION]/__init__.py \
  --config mage_integrations/TEST_CONFIG_S.json \
  --catalog mage_integrations/TEST_CATALOG.json \
  --state mage_integrations/TEST_STATE.json > mage_integrations/TEST_OUTPUT

Check TEST_OUTPUT to see real-time results!

Test writing your output to a database

Now, let's test writing our output to a destination. Populate the destination config file with a sample configuration in a similar manner to the source config, then run:

python3 mage_integrations/destinations/postgresql/__init__.py \
  --config mage_integrations/TEST_CONFIG_D.json \
  --state mage_integrations/TEST_STATE.json \
  --input_file_path mage_integrations/TEST_OUTPUT \
  --debug

To write TEST_OUTPUT to your destination. Note: you will need a sample data source to write to.

Source to destination end-to-end

This will test pulling from the target and writing to the destination:

python3 mage_integrations/sources/[SOURCE_INTEGRATION]/__init__.py \
  --config mage_integrations/TEST_CONFIG_S.json \
  --catalog mage_integrations/TEST_CATALOG.json \
  --state mage_integrations/TEST_STATE.json | python3 mage_integrations/destinations/[TARGET_INTEGRATION]/__init__.py \
  --config mage_integrations/TEST_CONFIG_D.json \
  --state mage_integrations/TEST_STATE.json \
  --debug

For example, an end-to-end GitHub to Postgres data integration:

python3 mage_integrations/sources/github/__init__.py \
  --config mage_integrations/TEST_CONFIG_S.json \
  --catalog mage_integrations/TEST_CATALOG.json \
  --state mage_integrations/TEST_STATE.json | python3 mage_integrations/destinations/postgres/__init__.py \
  --config mage_integrations/TEST_CONFIG_D.json \
  --state mage_integrations/TEST_STATE.json \
  --debug

Test in the Mage UI

Once you've tested your tap in the terminal, it's time to test it in Mage.

First, return to your terminal and run pip install -U mage_integrations/ in your mage_integrations directory. That will build our new mage-integrations package and make the changes you made available to the UI.

Open up Mage (localhost:3000 in dev) and create a new data integration pipeline:

Select your source from the list:

Now, perform the following in the Mage UI to verify a working source:

  • Test the connection
  • View and select streams
  • Sync one stream to a destination
    • If you're adding a tap in a PR, be sure to add logs of the source and show data in the destination table to the PR description.
  • If incremental sync is supported, please also test it: check if the state is updated and fetched correctly.

Count records

You can count the number of records in your stream with the following command:

python3 mage_integrations/sources/[INTEGRATION]/__init__.py \
  --config mage_integrations/TEST_CONFIG_S.json \
  --catalog mage_integrations/TEST_CATALOG.json \
  --state mage_integrations/TEST_STATE.json \
  --count_records \
  --selected_streams '["your_stream"]'

Perform a sample query

Use this template to perform a sample query of your data:

python3 mage_integrations/sources/freshdesk/__init__.py \
  --config mage_integrations/TEST_CONFIG_S.json \
  --catalog mage_integrations/TEST_CATALOG.json \
  --query_json '{"_end_date": null, "_execution_date": "2022-11-17T21:05:53.341319", "_execution_partition": "444/20221117T210443", "_start_date": null, "_limit": 1000, "_offset": 0}' \
  --state mage_integrations/TEST_STATE.json