Backport PR #551 on branch 1.x (Upgrades openai to version 1, removes…

… openai history in magics)" (#567) * Base chat handler refactor for custom slash commands (#398) * Adds attributes, starts adding to subclasses * Consistent syntax * Help for all handlers * Fix slash ID error * Iterate through entry points * Fix typo in call to select() * Moves config to magics, modifies extensions to attempt to load classes * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Moves config to proper location, improves error logging * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * WIP: Updates per feedback, adds custom handler * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Removes redundant code, style fixes * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Removes unnecessary custom message * Instantiates class * Validates slash ID * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Consistent arguments to chat handlers * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Refactors to avoid intentionally unused params * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Updates docs, removes custom handler from source and config * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Renames process_message to match base class * Adds needed parameter that had been deleted * Joins lines in contributor doc * Removes natural language routing type, which is not yet used * Update docs/source/developers/index.md Co-authored-by: Piyush Jain <[email protected]> * Update docs/source/developers/index.md Co-authored-by: Piyush Jain <[email protected]> * Update docs/source/developers/index.md Co-authored-by: Piyush Jain <[email protected]> * Revises per @3coins, avoids Latinism * Removes Configurable, since we do not yet have configurable traits * Uses Literal for validation * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci --------- Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-by: Piyush Jain <[email protected]> (cherry picked from commit 64e6daf) * Backport PR #551: Upgrades openai to version 1, removes openai history in magics
jupyterlab · Jan 5, 2024 · 8d3f22c · 8d3f22c
1 parent d91e9bb
commit 8d3f22c
Show file tree

Hide file tree

Showing 11 changed files with 28 additions and 181 deletions.
diff --git a/README.md b/README.md
@@ -31,7 +31,7 @@ In addition, you will need access to at least one model provider.
 
 To use any AI model provider within this notebook, you'll need the appropriate credentials, such as API keys.
 
-Obtain the necessary credentials (e.g., API keys) from your model provider's platform.
+Obtain the necessary credentials, such as API keys, from your model provider's platform.
 
 You can set your keys using environment variables or in a code cell in your notebook.
 In a code cell, you can use the %env magic command to set the credentials as follows:

diff --git a/docs/source/users/index.md b/docs/source/users/index.md
@@ -259,7 +259,7 @@ should contain the following data:
 AWS Console at the URL
 `https://<region>.console.aws.amazon.com/sagemaker/home?region=<region>#/endpoints`.
 
-- **Region name**: The AWS region your SageMaker endpoint is hosted in, e.g. `us-west-2`.
+- **Region name**: The AWS region your SageMaker endpoint is hosted in, such as `us-west-2`.
 
 - **Request schema**: The JSON object the endpoint expects, with the prompt
 being substituted into any value that matches the string literal `"<prompt>"`.
@@ -473,8 +473,8 @@ running the following code in a notebook cell or IPython shell:
 This command should not produce any output.
 
 :::{note}
-If you are using remote kernels (e.g.  Amazon SageMaker Studio), the above
-command will throw an error. This means that need to install the magics package
+If you are using remote kernels, such as in Amazon SageMaker Studio, the above
+command will throw an error. You will need to install the magics package
 on your remote kernel separately, even if you already have `jupyter_ai_magics`
 installed in your server's environment. In a notebook, run
 
@@ -519,11 +519,6 @@ We currently support the following language model providers:
 - `openai-chat`
 - `sagemaker-endpoint`
 
-:::{warning}
-As of v0.8.0, only the `%%ai` *cell* magic may be used to invoke a language
-model, while the `%ai` *line* magic is reserved for invoking subcommands.
-:::
-
 ### Listing available models
 
 Jupyter AI also includes multiple subcommands, which may be invoked via the
@@ -604,21 +599,6 @@ A function that computes the lowest common multiples of two integers, and
 a function that runs 5 test cases of the lowest common multiple function
 ```
 
-### Clearing the OpenAI chat history
-
-With the `openai-chat` provider *only*, you can run a cell magic command using the `-r` or
-`--reset` option to clear the chat history. After you do this, previous magic commands you've
-run with the `openai-chat` provider will no longer be added as context in
-requests to this provider.
-
-Because the `%%ai` command is a cell magic, you must provide a prompt on the second line.
-This prompt will not be sent to the provider. A reset command will not generate any output.
-
-```
-%%ai openai-chat:gpt-3.5-turbo -r
-reset the chat history
-```
-
 ### Interpolating in prompts
 
 Using curly brace syntax, you can include variables and other Python expressions in your

diff --git a/examples/commands.ipynb b/examples/commands.ipynb
diff --git a/packages/jupyter-ai-magics/jupyter_ai_magics/__init__.py b/packages/jupyter-ai-magics/jupyter_ai_magics/__init__.py
@@ -22,7 +22,6 @@
     BedrockChatProvider,
     BedrockProvider,
     ChatAnthropicProvider,
-    ChatOpenAINewProvider,
     ChatOpenAIProvider,
     CohereProvider,
     GPT4AllProvider,

diff --git a/packages/jupyter-ai-magics/jupyter_ai_magics/magics.py b/packages/jupyter-ai-magics/jupyter_ai_magics/magics.py
@@ -137,20 +137,22 @@ def __init__(self, shell):
         super().__init__(shell)
         self.transcript_openai = []
 
-        # suppress warning when using old OpenAIChat provider
-        warnings.filterwarnings(
-            "ignore",
-            message="You are trying to use a chat model. This way of initializing it is "
-            "no longer supported. Instead, please use: "
-            "`from langchain.chat_models import ChatOpenAI`",
-        )
         # suppress warning when using old Anthropic provider
         warnings.filterwarnings(
             "ignore",
             message="This Anthropic LLM is deprecated. Please use "
             "`from langchain.chat_models import ChatAnthropic` instead",
         )
 
+        # suppress warning about our exception handler
+        warnings.filterwarnings(
+            "ignore",
+            message="IPython detected, but you already "
+            "have a custom exception handler installed. I'll skip installing "
+            "Trio's custom handler, but this means exception groups will not "
+            "show full tracebacks.",
+        )
+
         self.providers = get_lm_providers()
 
         # initialize a registry of custom model/chain names
@@ -421,17 +423,10 @@ def handle_error(self, args: ErrorArgs):
         # Set CellArgs based on ErrorArgs
         values = args.dict()
         values["type"] = "root"
-        values["reset"] = False
         cell_args = CellArgs(**values)
 
         return self.run_ai_cell(cell_args, prompt)
 
-    def _append_exchange_openai(self, prompt: str, output: str):
-        """Appends a conversational exchange between user and an OpenAI Chat
-        model to a transcript that will be included in future exchanges."""
-        self.transcript_openai.append({"role": "user", "content": prompt})
-        self.transcript_openai.append({"role": "assistant", "content": output})
-
     def _decompose_model_id(self, model_id: str):
         """Breaks down a model ID into a two-tuple (provider_id, local_model_id). Returns (None, None) if indeterminate."""
         # custom_model_registry maps keys to either a model name (a string) or an LLMChain.
@@ -534,11 +529,6 @@ def run_ai_cell(self, args: CellArgs, prompt: str):
                 + "If you were trying to run a command, run `%ai help` to see a list of commands.",
             )
 
-        # if `--reset` is specified, reset transcript and return early
-        if provider_id == "openai-chat" and args.reset:
-            self.transcript_openai = []
-            return
-
         # validate presence of authn credentials
         auth_strategy = self.providers[provider_id].auth_strategy
         if auth_strategy:
@@ -561,8 +551,6 @@ def run_ai_cell(self, args: CellArgs, prompt: str):
 
         # configure and instantiate provider
         provider_params = {"model_id": local_model_id}
-        if provider_id == "openai-chat":
-            provider_params["prefix_messages"] = self.transcript_openai
         # for SageMaker, validate that required params are specified
         if provider_id == "sagemaker-endpoint":
             if (
@@ -596,11 +584,6 @@ def run_ai_cell(self, args: CellArgs, prompt: str):
             result = provider.generate([prompt])
 
         output = result.generations[0][0].text
-
-        # if openai-chat, append exchange to transcript
-        if provider_id == "openai-chat":
-            self._append_exchange_openai(prompt, output)
-
         md = {"jupyter_ai": {"provider_id": provider_id, "model_id": local_model_id}}
 
         return self.display_output(output, args.format, md)

diff --git a/packages/jupyter-ai-magics/jupyter_ai_magics/parsers.py b/packages/jupyter-ai-magics/jupyter_ai_magics/parsers.py
@@ -46,15 +46,14 @@ class CellArgs(BaseModel):
     type: Literal["root"] = "root"
     model_id: str
     format: FORMAT_CHOICES_TYPE
-    reset: bool
     model_parameters: Optional[str]
     # The following parameters are required only for SageMaker models
     region_name: Optional[str]
     request_schema: Optional[str]
     response_path: Optional[str]
 
 
-# Should match CellArgs, but without "reset"
+# Should match CellArgs
 class ErrorArgs(BaseModel):
     type: Literal["error"] = "error"
     model_id: str
@@ -126,13 +125,6 @@ def verify_json_value(ctx, param, value):
     default="markdown",
     help=FORMAT_HELP,
 )
-@click.option(
-    "-r",
-    "--reset",
-    is_flag=True,
-    help="""Clears the conversation transcript used when interacting with an
-    OpenAI chat model provider. Does nothing with other providers.""",
-)
 @click.option(
     REGION_NAME_SHORT_OPTION,
     REGION_NAME_LONG_OPTION,

diff --git a/packages/jupyter-ai-magics/jupyter_ai_magics/providers.py b/packages/jupyter-ai-magics/jupyter_ai_magics/providers.py
@@ -22,7 +22,6 @@
     AzureChatOpenAI,
     BedrockChat,
     ChatAnthropic,
-    ChatOpenAI,
     QianfanChatEndpoint,
 )
 from langchain.chat_models.base import BaseChatModel
@@ -44,6 +43,7 @@
 from langchain.pydantic_v1 import BaseModel, Extra, root_validator
 from langchain.schema import LLMResult
 from langchain.utils import get_from_dict_or_env
+from langchain_community.chat_models import ChatOpenAI
 
 
 class EnvAuthStrategy(BaseModel):
@@ -534,13 +534,13 @@ def is_api_key_exc(cls, e: Exception):
         """
         import openai
 
-        if isinstance(e, openai.error.AuthenticationError):
+        if isinstance(e, openai.AuthenticationError):
             error_details = e.json_body.get("error", {})
             return error_details.get("code") == "invalid_api_key"
         return False
 
 
-class ChatOpenAIProvider(BaseProvider, OpenAIChat):
+class ChatOpenAIProvider(BaseProvider, ChatOpenAI):
     id = "openai-chat"
     name = "OpenAI"
     models = [
@@ -561,48 +561,6 @@ class ChatOpenAIProvider(BaseProvider, OpenAIChat):
     pypi_package_deps = ["openai"]
     auth_strategy = EnvAuthStrategy(name="OPENAI_API_KEY")
 
-    def append_exchange(self, prompt: str, output: str):
-        """Appends a conversational exchange between user and an OpenAI Chat
-        model to a transcript that will be included in future exchanges."""
-        self.prefix_messages.append({"role": "user", "content": prompt})
-        self.prefix_messages.append({"role": "assistant", "content": output})
-
-    @classmethod
-    def is_api_key_exc(cls, e: Exception):
-        """
-        Determine if the exception is an OpenAI API key error.
-        """
-        import openai
-
-        if isinstance(e, openai.error.AuthenticationError):
-            error_details = e.json_body.get("error", {})
-            return error_details.get("code") == "invalid_api_key"
-        return False
-
-
-# uses the new OpenAIChat provider. temporarily living as a separate class until
-# conflicts can be resolved
-class ChatOpenAINewProvider(BaseProvider, ChatOpenAI):
-    id = "openai-chat-new"
-    name = "OpenAI"
-    models = [
-        "gpt-3.5-turbo",
-        "gpt-3.5-turbo-16k",
-        "gpt-3.5-turbo-0301",
-        "gpt-3.5-turbo-0613",
-        "gpt-3.5-turbo-16k-0613",
-        "gpt-4",
-        "gpt-4-0314",
-        "gpt-4-0613",
-        "gpt-4-32k",
-        "gpt-4-32k-0314",
-        "gpt-4-32k-0613",
-        "gpt-4-1106-preview",
-    ]
-    model_id_key = "model_name"
-    pypi_package_deps = ["openai"]
-    auth_strategy = EnvAuthStrategy(name="OPENAI_API_KEY")
-
     fields = [
         TextField(
             key="openai_api_base", label="Base API URL (optional)", format="text"
@@ -620,7 +578,7 @@ def is_api_key_exc(cls, e: Exception):
         """
         import openai
 
-        if isinstance(e, openai.error.AuthenticationError):
+        if isinstance(e, openai.AuthenticationError):
             error_details = e.json_body.get("error", {})
             return error_details.get("code") == "invalid_api_key"
         return False

diff --git a/packages/jupyter-ai-magics/pyproject.toml b/packages/jupyter-ai-magics/pyproject.toml
@@ -50,7 +50,7 @@ all = [
     "huggingface_hub",
     "ipywidgets",
     "pillow",
-    "openai",
+    "openai~=1.6.1",
     "boto3",
     "qianfan"
 ]
@@ -63,7 +63,6 @@ gpt4all = "jupyter_ai_magics:GPT4AllProvider"
 huggingface_hub = "jupyter_ai_magics:HfHubProvider"
 openai = "jupyter_ai_magics:OpenAIProvider"
 openai-chat = "jupyter_ai_magics:ChatOpenAIProvider"
-openai-chat-new = "jupyter_ai_magics:ChatOpenAINewProvider"
 azure-chat-openai = "jupyter_ai_magics:AzureChatOpenAIProvider"
 sagemaker-endpoint = "jupyter_ai_magics:SmEndpointProvider"
 amazon-bedrock = "jupyter_ai_magics:BedrockProvider"

diff --git a/packages/jupyter-ai/jupyter_ai/handlers.py b/packages/jupyter-ai/jupyter_ai/handlers.py
@@ -294,10 +294,6 @@ def get(self):
 
         # Step 1: gather providers
         for provider in self.lm_providers.values():
-            # skip old legacy OpenAI chat provider used only in magics
-            if provider.id == "openai-chat":
-                continue
-
             optionals = {}
             if provider.model_id_label:
                 optionals["model_id_label"] = provider.model_id_label

diff --git a/packages/jupyter-ai/jupyter_ai/tests/test_config_manager.py b/packages/jupyter-ai/jupyter_ai/tests/test_config_manager.py
@@ -127,7 +127,7 @@ def configure_to_cohere(cm: ConfigManager):
 def configure_to_openai(cm: ConfigManager):
     """Configures the ConfigManager to use OpenAI language and embedding models
     with the API key set. Returns a 3-tuple of the keyword arguments used."""
-    LM_GID = "openai-chat-new:gpt-3.5-turbo"
+    LM_GID = "openai-chat:gpt-3.5-turbo"
     EM_GID = "openai:text-embedding-ada-002"
     API_KEYS = {"OPENAI_API_KEY": "foobar"}
     LM_LID = "gpt-3.5-turbo"
@@ -157,7 +157,7 @@ def test_init_with_blocklists(cm: ConfigManager, common_cm_kwargs):
     del cm
 
     blocked_providers = ["openai"]  # blocks EM
-    blocked_models = ["openai-chat-new:gpt-3.5-turbo"]  # blocks LM
+    blocked_models = ["openai-chat:gpt-3.5-turbo"]  # blocks LM
     kwargs = {
         **common_cm_kwargs,
         "blocked_providers": blocked_providers,
@@ -278,14 +278,14 @@ def test_forbid_write_write_conflict(cm: ConfigManager):
 
     # call UpdateConfig separately after DescribeConfig with `last_read` unset
     # to force a write
-    cm.update_config(UpdateConfigRequest(model_provider_id="openai-chat-new:gpt-4"))
+    cm.update_config(UpdateConfigRequest(model_provider_id="openai-chat:gpt-4"))
 
     # this update should fail, as this generates a write-write conflict (where
     # the second update clobbers the first update).
     with pytest.raises(WriteConflictError):
         cm.update_config(
             UpdateConfigRequest(
-                model_provider_id="openai-chat-new:gpt-4-32k", last_read=last_read
+                model_provider_id="openai-chat:gpt-4-32k", last_read=last_read
             )
         )
 

diff --git a/packages/jupyter-ai/pyproject.toml b/packages/jupyter-ai/pyproject.toml
@@ -24,7 +24,6 @@ classifiers = [
 dependencies = [
     "jupyter_server>=1.6,<3",
     "jupyterlab>=3.5,<4",
-    "openai~=0.26",
     "aiosqlite>=0.18",
     "importlib_metadata>=5.2.0",
     "langchain==0.0.350",