From 5b00885b4917fddfea2420895d3f144485e8aa43 Mon Sep 17 00:00:00 2001 From: Dobiichi-Origami <56953648+Dobiichi-Origami@users.noreply.github.com> Date: Wed, 8 May 2024 23:35:10 +0800 Subject: [PATCH] community: add `bind_tools` and `with_structured_output` support to `QianfanChatEndpoint` (#21412) MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit …Endpoint` Thank you for contributing to LangChain! - [x] **PR title**: "package: description" - Where "package" is whichever of langchain, community, core, experimental, etc. is being modified. Use "docs: ..." for purely docs changes, "templates: ..." for template changes, "infra: ..." for CI changes. - Example: "community: add foobar LLM" - [x] **PR message**: ***Delete this entire checklist*** and replace with - **Description:** add `bind_tools` and `with_structured_output` support to `QianfanChatEndpoint` - [x] **Lint and test**: Run `make format`, `make lint` and `make test` from the root of the package(s) you've modified. See contribution guidelines for more: https://python.langchain.com/docs/contributing/ --- .../chat_models/baidu_qianfan_endpoint.py | 184 +++++++++++++++++- 1 file changed, 182 insertions(+), 2 deletions(-) diff --git a/libs/community/langchain_community/chat_models/baidu_qianfan_endpoint.py b/libs/community/langchain_community/chat_models/baidu_qianfan_endpoint.py index d39666df559..f25f44c933e 100644 --- a/libs/community/langchain_community/chat_models/baidu_qianfan_endpoint.py +++ b/libs/community/langchain_community/chat_models/baidu_qianfan_endpoint.py @@ -1,10 +1,25 @@ import logging -from typing import Any, AsyncIterator, Dict, Iterator, List, Mapping, Optional, cast +from operator import itemgetter +from typing import ( + Any, + AsyncIterator, + Callable, + Dict, + Iterator, + List, + Mapping, + Optional, + Sequence, + Type, + Union, + cast, +) from langchain_core.callbacks import ( AsyncCallbackManagerForLLMRun, CallbackManagerForLLMRun, ) +from langchain_core.language_models import LanguageModelInput from langchain_core.language_models.chat_models import BaseChatModel from langchain_core.messages import ( AIMessage, @@ -15,9 +30,17 @@ from langchain_core.messages import ( HumanMessage, SystemMessage, ) +from langchain_core.output_parsers.base import OutputParserLike +from langchain_core.output_parsers.openai_tools import ( + JsonOutputKeyToolsParser, + PydanticToolsParser, +) from langchain_core.outputs import ChatGeneration, ChatGenerationChunk, ChatResult -from langchain_core.pydantic_v1 import Field, SecretStr, root_validator +from langchain_core.pydantic_v1 import BaseModel, Field, SecretStr, root_validator +from langchain_core.runnables import Runnable, RunnableMap, RunnablePassthrough +from langchain_core.tools import BaseTool from langchain_core.utils import convert_to_secret_str, get_from_dict_or_env +from langchain_core.utils.function_calling import convert_to_openai_tool logger = logging.getLogger(__name__) @@ -66,6 +89,12 @@ def _convert_dict_to_message(_dict: Mapping[str, Any]) -> AIMessage: object=additional_kwargs.get("object", ""), search_info=additional_kwargs.get("search_info", []), function_call=additional_kwargs.get("function_call", {}), + tool_calls=[ + { + "type": "function", + "function": additional_kwargs.get("function_call", {}), + } + ], ), ) @@ -391,3 +420,154 @@ class QianfanChatEndpoint(BaseChatModel): if run_manager: await run_manager.on_llm_new_token(chunk.text, chunk=chunk) yield chunk + + def bind_tools( + self, + tools: Sequence[Union[Dict[str, Any], Type[BaseModel], Callable, BaseTool]], + **kwargs: Any, + ) -> Runnable[LanguageModelInput, BaseMessage]: + """Bind tool-like objects to this chat model. + + Assumes model is compatible with OpenAI tool-calling API. + + Args: + tools: A list of tool definitions to bind to this chat model. + Can be a dictionary, pydantic model, callable, or BaseTool. Pydantic + models, callables, and BaseTools will be automatically converted to + their schema dictionary representation. + **kwargs: Any additional parameters to pass to the + :class:`~langchain.runnable.Runnable` constructor. + """ + + formatted_tools = [convert_to_openai_tool(tool)["function"] for tool in tools] + return super().bind(functions=formatted_tools, **kwargs) + + def with_structured_output( + self, + schema: Union[Dict, Type[BaseModel]], + *, + include_raw: bool = False, + **kwargs: Any, + ) -> Runnable[LanguageModelInput, Union[Dict, BaseModel]]: + """Model wrapper that returns outputs formatted to match the given schema. + + Args: + schema: The output schema as a dict or a Pydantic class. If a Pydantic class + then the model output will be an object of that class. If a dict then + the model output will be a dict. With a Pydantic class the returned + attributes will be validated, whereas with a dict they will not be. If + `method` is "function_calling" and `schema` is a dict, then the dict + must match the OpenAI function-calling spec. + include_raw: If False then only the parsed structured output is returned. If + an error occurs during model output parsing it will be raised. If True + then both the raw model response (a BaseMessage) and the parsed model + response will be returned. If an error occurs during output parsing it + will be caught and returned as well. The final output is always a dict + with keys "raw", "parsed", and "parsing_error". + + Returns: + A Runnable that takes any ChatModel input and returns as output: + + If include_raw is True then a dict with keys: + raw: BaseMessage + parsed: Optional[_DictOrPydantic] + parsing_error: Optional[BaseException] + + If include_raw is False then just _DictOrPydantic is returned, + where _DictOrPydantic depends on the schema: + + If schema is a Pydantic class then _DictOrPydantic is the Pydantic + class. + + If schema is a dict then _DictOrPydantic is a dict. + + Example: Function-calling, Pydantic schema (method="function_calling", include_raw=False): + .. code-block:: python + + from langchain_mistralai import QianfanChatEndpoint + from langchain_core.pydantic_v1 import BaseModel + + class AnswerWithJustification(BaseModel): + '''An answer to the user question along with justification for the answer.''' + answer: str + justification: str + + llm = QianfanChatEndpoint(endpoint="ernie-3.5-8k-0329") + structured_llm = llm.with_structured_output(AnswerWithJustification) + + structured_llm.invoke("What weighs more a pound of bricks or a pound of feathers") + + # -> AnswerWithJustification( + # answer='They weigh the same', + # justification='Both a pound of bricks and a pound of feathers weigh one pound. The weight is the same, but the volume or density of the objects may differ.' + # ) + + Example: Function-calling, Pydantic schema (method="function_calling", include_raw=True): + .. code-block:: python + + from langchain_mistralai import QianfanChatEndpoint + from langchain_core.pydantic_v1 import BaseModel + + class AnswerWithJustification(BaseModel): + '''An answer to the user question along with justification for the answer.''' + answer: str + justification: str + + llm = QianfanChatEndpoint(endpoint="ernie-3.5-8k-0329") + structured_llm = llm.with_structured_output(AnswerWithJustification, include_raw=True) + + structured_llm.invoke("What weighs more a pound of bricks or a pound of feathers") + # -> { + # 'raw': AIMessage(content='', additional_kwargs={'tool_calls': [{'id': 'call_Ao02pnFYXD6GN1yzc0uXPsvF', 'function': {'arguments': '{"answer":"They weigh the same.","justification":"Both a pound of bricks and a pound of feathers weigh one pound. The weight is the same, but the volume or density of the objects may differ."}', 'name': 'AnswerWithJustification'}, 'type': 'function'}]}), + # 'parsed': AnswerWithJustification(answer='They weigh the same.', justification='Both a pound of bricks and a pound of feathers weigh one pound. The weight is the same, but the volume or density of the objects may differ.'), + # 'parsing_error': None + # } + + Example: Function-calling, dict schema (method="function_calling", include_raw=False): + .. code-block:: python + + from langchain_mistralai import QianfanChatEndpoint + from langchain_core.pydantic_v1 import BaseModel + from langchain_core.utils.function_calling import convert_to_openai_tool + + class AnswerWithJustification(BaseModel): + '''An answer to the user question along with justification for the answer.''' + answer: str + justification: str + + dict_schema = convert_to_openai_tool(AnswerWithJustification) + llm = QianfanChatEndpoint(endpoint="ernie-3.5-8k-0329") + structured_llm = llm.with_structured_output(dict_schema) + + structured_llm.invoke("What weighs more a pound of bricks or a pound of feathers") + # -> { + # 'answer': 'They weigh the same', + # 'justification': 'Both a pound of bricks and a pound of feathers weigh one pound. The weight is the same, but the volume and density of the two substances differ.' + # } + + """ # noqa: E501 + if kwargs: + raise ValueError(f"Received unsupported arguments {kwargs}") + is_pydantic_schema = isinstance(schema, type) and issubclass(schema, BaseModel) + llm = self.bind_tools([schema]) + if is_pydantic_schema: + output_parser: OutputParserLike = PydanticToolsParser( + tools=[schema], first_tool_only=True + ) + else: + key_name = convert_to_openai_tool(schema)["function"]["name"] + output_parser = JsonOutputKeyToolsParser( + key_name=key_name, first_tool_only=True + ) + + if include_raw: + parser_assign = RunnablePassthrough.assign( + parsed=itemgetter("raw") | output_parser, parsing_error=lambda _: None + ) + parser_none = RunnablePassthrough.assign(parsed=lambda _: None) + parser_with_fallback = parser_assign.with_fallbacks( + [parser_none], exception_key="parsing_error" + ) + return RunnableMap(raw=llm) | parser_with_fallback + else: + return llm | output_parser