github/langchain
1
0
Fork 0
mirror of https://github.com/hwchase17/langchain.git synced 2025-09-25 04:49:17 +00:00
Code Issues Packages Projects Releases Wiki Activity

core[patch], integrations[patch]: convert TypedDict to tool schema support (#24641)

Browse Source 
supports following UX

```python
    class SubTool(TypedDict):
        """Subtool docstring"""

        args: Annotated[Dict[str, Any], {}, "this does bar"]

    class Tool(TypedDict):
        """Docstring
        Args:
            arg1: foo
        """

        arg1: str
        arg2: Union[int, str]
        arg3: Optional[List[SubTool]]
        arg4: Annotated[Literal["bar", "baz"], ..., "this does foo"]
        arg5: Annotated[Optional[float], None]
```

- can parse google style docstring
- can use Annotated to specify default value (second arg)
- can use Annotated to specify arg description (third arg)
- can have nested complex types
This commit is contained in:
Bagatur
2024-07-31 11:27:24 -07:00
committed by GitHub
parent d24b82357f
commit 8461934c2b
17 changed files with 1371 additions and 468 deletions
Download Patch File Download Diff File Expand all files Collapse all files

121
libs/partners/openai/langchain_openai/chat_models/azure.py
View File

@@ -652,7 +652,7 @@ class AzureChatOpenAI(BaseChatOpenAI):
def bind_tools(
self,
tools: Sequence[Union[Dict[str, Any], Type[BaseModel], Callable, BaseTool]],
tools: Sequence[Union[Dict[str, Any], Type, Callable, BaseTool]],
*,
tool_choice: Optional[
Union[dict, str, Literal["auto", "none", "required", "any"], bool]
@@ -703,20 +703,27 @@ class AzureChatOpenAI(BaseChatOpenAI):
"""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 or be a valid JSON schema
with top level 'title' and 'description' keys specified.
method: The method for steering model generation, either "function_calling"
schema:
The output schema. Can be passed in as:
- an OpenAI function/tool schema,
- a JSON Schema,
- a TypedDict class,
- or a Pydantic class.
If ``schema`` is a Pydantic class then the model output will be a
Pydantic instance of that class, and the model-generated fields will be
validated by the Pydantic class. Otherwise the model output will be a
dict and will not be validated. See :meth:`langchain_core.utils.function_calling.convert_to_openai_tool`
for more on how to properly specify types and descriptions of
schema fields when specifying a Pydantic or TypedDict class.
method:
The method for steering model generation, either "function_calling"
or "json_mode". If "function_calling" then the schema will be converted
to an OpenAI function and the returned model will make use of the
function-calling API. If "json_mode" then OpenAI's JSON mode will be
used. Note that if using "json_mode" then you must include instructions
for formatting the output into the desired schema into the model call.
include_raw: If False then only the parsed structured output is returned. If
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
@@ -724,36 +731,40 @@ class AzureChatOpenAI(BaseChatOpenAI):
with keys "raw", "parsed", and "parsing_error".
Returns:
A Runnable that takes any ChatModel input and returns as output:
A Runnable that takes same inputs as a :class:`langchain_core.language_models.chat.BaseChatModel`.
If include_raw is True then a dict with keys:
raw: BaseMessage
parsed: Optional[_DictOrPydantic]
parsing_error: Optional[BaseException]
If ``include_raw`` is False and ``schema`` is a Pydantic class, Runnable outputs
an instance of ``schema`` (i.e., a Pydantic object).
If include_raw is False then just _DictOrPydantic is returned,
where _DictOrPydantic depends on the schema:
Otherwise, if ``include_raw`` is False then Runnable outputs a dict.
If schema is a Pydantic class then _DictOrPydantic is the Pydantic
class.
If ``include_raw`` is True, then Runnable outputs a dict with keys:
- ``"raw"``: BaseMessage
- ``"parsed"``: None if there was a parsing error, otherwise the type depends on the ``schema`` as described above.
- ``"parsing_error"``: Optional[BaseException]
If schema is a dict then _DictOrPydantic is a dict.
Example: Function-calling, Pydantic schema (method="function_calling", include_raw=False):
Example: schema=Pydantic class, method="function_calling", include_raw=False:
.. code-block:: python
from typing import Optional
from langchain_openai import AzureChatOpenAI
from langchain_core.pydantic_v1 import BaseModel
from langchain_core.pydantic_v1 import BaseModel, Field
class AnswerWithJustification(BaseModel):
'''An answer to the user question along with justification for the answer.'''
answer: str
justification: str
# If we provide default values and/or descriptions for fields, these will be passed
# to the model. This is an important part of improving a model's ability to
# correctly return structured outputs.
justification: Optional[str] = Field(
default=None, description="A justification for the answer."
)
llm = AzureChatOpenAI(azure_deployment="gpt-35-turbo", temperature=0)
llm = AzureChatOpenAI(model="gpt-3.5-turbo-0125", temperature=0)
structured_llm = llm.with_structured_output(AnswerWithJustification)
structured_llm.invoke(
@@ -765,7 +776,7 @@ class AzureChatOpenAI(BaseChatOpenAI):
# 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):
Example: schema=Pydantic class, method="function_calling", include_raw=True:
.. code-block:: python
from langchain_openai import AzureChatOpenAI
@@ -779,7 +790,7 @@ class AzureChatOpenAI(BaseChatOpenAI):
justification: str
llm = AzureChatOpenAI(azure_deployment="gpt-35-turbo", temperature=0)
llm = AzureChatOpenAI(model="gpt-3.5-turbo-0125", temperature=0)
structured_llm = llm.with_structured_output(
AnswerWithJustification, include_raw=True
)
@@ -793,24 +804,27 @@ class AzureChatOpenAI(BaseChatOpenAI):
# 'parsing_error': None
# }
Example: Function-calling, dict schema (method="function_calling", include_raw=False):
Example: schema=TypedDict class, method="function_calling", include_raw=False:
.. code-block:: python
# IMPORTANT: If you are using Python <=3.8, you need to import Annotated
# from typing_extensions, not from typing.
from typing_extensions import Annotated, TypedDict
from langchain_openai import AzureChatOpenAI
from langchain_core.pydantic_v1 import BaseModel
from langchain_core.utils.function_calling import convert_to_openai_tool
class AnswerWithJustification(BaseModel):
class AnswerWithJustification(TypedDict):
'''An answer to the user question along with justification for the answer.'''
answer: str
justification: str
justification: Annotated[
Optional[str], None, "A justification for the answer."
]
dict_schema = convert_to_openai_tool(AnswerWithJustification)
llm = AzureChatOpenAI(azure_deployment="gpt-35-turbo", temperature=0)
structured_llm = llm.with_structured_output(dict_schema)
llm = AzureChatOpenAI(model="gpt-3.5-turbo-0125", temperature=0)
structured_llm = llm.with_structured_output(AnswerWithJustification)
structured_llm.invoke(
"What weighs more a pound of bricks or a pound of feathers"
@@ -820,7 +834,36 @@ class AzureChatOpenAI(BaseChatOpenAI):
# '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.'
# }
Example: JSON mode, Pydantic schema (method="json_mode", include_raw=True):
Example: schema=OpenAI function schema, method="function_calling", include_raw=False:
.. code-block:: python
from langchain_openai import AzureChatOpenAI
oai_schema = {
'name': 'AnswerWithJustification',
'description': 'An answer to the user question along with justification for the answer.',
'parameters': {
'type': 'object',
'properties': {
'answer': {'type': 'string'},
'justification': {'description': 'A justification for the answer.', 'type': 'string'}
},
'required': ['answer']
}
}
llm = AzureChatOpenAI(model="gpt-3.5-turbo-0125", temperature=0)
structured_llm = llm.with_structured_output(oai_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.'
# }
Example: schema=Pydantic class, method="json_mode", include_raw=True:
.. code-block::
from langchain_openai import AzureChatOpenAI
@@ -830,7 +873,7 @@ class AzureChatOpenAI(BaseChatOpenAI):
answer: str
justification: str
llm = AzureChatOpenAI(azure_deployment="gpt-35-turbo", temperature=0)
llm = AzureChatOpenAI(model="gpt-3.5-turbo-0125", temperature=0)
structured_llm = llm.with_structured_output(
AnswerWithJustification,
method="json_mode",
@@ -848,7 +891,7 @@ class AzureChatOpenAI(BaseChatOpenAI):
# 'parsing_error': None
# }
Example: JSON mode, no schema (schema=None, method="json_mode", include_raw=True):
Example: schema=None, method="json_mode", include_raw=True:
.. code-block::
structured_llm = llm.with_structured_output(method="json_mode", include_raw=True)
@@ -866,8 +909,6 @@ class AzureChatOpenAI(BaseChatOpenAI):
# },
# 'parsing_error': None
# }
""" # noqa: E501
if kwargs:
raise ValueError(f"Received unsupported arguments {kwargs}")

146
libs/partners/openai/langchain_openai/chat_models/base.py
View File

@@ -285,7 +285,7 @@ class _FunctionCall(TypedDict):
_BM = TypeVar("_BM", bound=BaseModel)
_DictOrPydanticClass = Union[Dict[str, Any], Type[_BM]]
_DictOrPydanticClass = Union[Dict[str, Any], Type[_BM], Type]
_DictOrPydantic = Union[Dict, _BM]
@@ -947,7 +947,7 @@ class BaseChatOpenAI(BaseChatModel):
def bind_tools(
self,
tools: Sequence[Union[Dict[str, Any], Type[BaseModel], Callable, BaseTool]],
tools: Sequence[Union[Dict[str, Any], Type, Callable, BaseTool]],
*,
tool_choice: Optional[
Union[dict, str, Literal["auto", "none", "required", "any"], bool]
@@ -960,23 +960,19 @@ class BaseChatOpenAI(BaseChatModel):
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.
Supports any tool definition handled by
:meth:`langchain_core.utils.function_calling.convert_to_openai_tool`.
tool_choice: Which tool to require the model to call.
Options are:
name of the tool (str): calls corresponding tool;
"auto": automatically selects a tool (including no tool);
"none": does not call a tool;
"any" or "required": force at least one tool to be called;
True: forces tool call (requires `tools` be length 1);
False: no effect;
or a dict of the form:
{"type": "function", "function": {"name": <<tool_name>>}}.
**kwargs: Any additional parameters to pass to the
:class:`~langchain.runnable.Runnable` constructor.
"""
- str of the form ``"<<tool_name>>"``: calls <<tool_name>> tool.
- ``"auto"``: automatically selects a tool (including no tool).
- ``"none"``: does not call a tool.
- ``"any"`` or ``"required"`` or ``True``: force at least one tool to be called.
- dict of the form ``{"type": "function", "function": {"name": <<tool_name>>}}``: calls <<tool_name>> tool.
- ``False`` or ``None``: no effect, default OpenAI behavior.
kwargs: Any additional parameters are passed directly to
``self.bind(**kwargs)``.
""" # noqa: E501
formatted_tools = [convert_to_openai_tool(tool) for tool in tools]
if tool_choice:
@@ -1046,20 +1042,32 @@ class BaseChatOpenAI(BaseChatModel):
"""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 or be a valid JSON schema
with top level 'title' and 'description' keys specified.
method: The method for steering model generation, either "function_calling"
schema:
The output schema. Can be passed in as:
- an OpenAI function/tool schema,
- a JSON Schema,
- a TypedDict class (support added in 0.1.20),
- or a Pydantic class.
If ``schema`` is a Pydantic class then the model output will be a
Pydantic instance of that class, and the model-generated fields will be
validated by the Pydantic class. Otherwise the model output will be a
dict and will not be validated. See :meth:`langchain_core.utils.function_calling.convert_to_openai_tool`
for more on how to properly specify types and descriptions of
schema fields when specifying a Pydantic or TypedDict class.
.. versionchanged:: 0.1.20
Added support for TypedDict class.
method:
The method for steering model generation, either "function_calling"
or "json_mode". If "function_calling" then the schema will be converted
to an OpenAI function and the returned model will make use of the
function-calling API. If "json_mode" then OpenAI's JSON mode will be
used. Note that if using "json_mode" then you must include instructions
for formatting the output into the desired schema into the model call.
include_raw: If False then only the parsed structured output is returned. If
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
@@ -1067,33 +1075,37 @@ class BaseChatOpenAI(BaseChatModel):
with keys "raw", "parsed", and "parsing_error".
Returns:
A Runnable that takes any ChatModel input and returns as output:
A Runnable that takes same inputs as a :class:`langchain_core.language_models.chat.BaseChatModel`.
If include_raw is True then a dict with keys:
raw: BaseMessage
parsed: Optional[_DictOrPydantic]
parsing_error: Optional[BaseException]
If ``include_raw`` is False and ``schema`` is a Pydantic class, Runnable outputs
an instance of ``schema`` (i.e., a Pydantic object).
If include_raw is False then just _DictOrPydantic is returned,
where _DictOrPydantic depends on the schema:
Otherwise, if ``include_raw`` is False then Runnable outputs a dict.
If schema is a Pydantic class then _DictOrPydantic is the Pydantic
class.
If ``include_raw`` is True, then Runnable outputs a dict with keys:
- ``"raw"``: BaseMessage
- ``"parsed"``: None if there was a parsing error, otherwise the type depends on the ``schema`` as described above.
- ``"parsing_error"``: Optional[BaseException]
If schema is a dict then _DictOrPydantic is a dict.
Example: Function-calling, Pydantic schema (method="function_calling", include_raw=False):
Example: schema=Pydantic class, method="function_calling", include_raw=False:
.. code-block:: python
from typing import Optional
from langchain_openai import ChatOpenAI
from langchain_core.pydantic_v1 import BaseModel
from langchain_core.pydantic_v1 import BaseModel, Field
class AnswerWithJustification(BaseModel):
'''An answer to the user question along with justification for the answer.'''
answer: str
justification: str
# If we provide default values and/or descriptions for fields, these will be passed
# to the model. This is an important part of improving a model's ability to
# correctly return structured outputs.
justification: Optional[str] = Field(
default=None, description="A justification for the answer."
)
llm = ChatOpenAI(model="gpt-3.5-turbo-0125", temperature=0)
@@ -1108,7 +1120,7 @@ class BaseChatOpenAI(BaseChatModel):
# 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):
Example: schema=Pydantic class, method="function_calling", include_raw=True:
.. code-block:: python
from langchain_openai import ChatOpenAI
@@ -1136,24 +1148,27 @@ class BaseChatOpenAI(BaseChatModel):
# 'parsing_error': None
# }
Example: Function-calling, dict schema (method="function_calling", include_raw=False):
Example: schema=TypedDict class, method="function_calling", include_raw=False:
.. code-block:: python
# IMPORTANT: If you are using Python <=3.8, you need to import Annotated
# from typing_extensions, not from typing.
from typing_extensions import Annotated, TypedDict
from langchain_openai import ChatOpenAI
from langchain_core.pydantic_v1 import BaseModel
from langchain_core.utils.function_calling import convert_to_openai_tool
class AnswerWithJustification(BaseModel):
class AnswerWithJustification(TypedDict):
'''An answer to the user question along with justification for the answer.'''
answer: str
justification: str
justification: Annotated[
Optional[str], None, "A justification for the answer."
]
dict_schema = convert_to_openai_tool(AnswerWithJustification)
llm = ChatOpenAI(model="gpt-3.5-turbo-0125", temperature=0)
structured_llm = llm.with_structured_output(dict_schema)
structured_llm = llm.with_structured_output(AnswerWithJustification)
structured_llm.invoke(
"What weighs more a pound of bricks or a pound of feathers"
@@ -1163,7 +1178,36 @@ class BaseChatOpenAI(BaseChatModel):
# '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.'
# }
Example: JSON mode, Pydantic schema (method="json_mode", include_raw=True):
Example: schema=OpenAI function schema, method="function_calling", include_raw=False:
.. code-block:: python
from langchain_openai import ChatOpenAI
oai_schema = {
'name': 'AnswerWithJustification',
'description': 'An answer to the user question along with justification for the answer.',
'parameters': {
'type': 'object',
'properties': {
'answer': {'type': 'string'},
'justification': {'description': 'A justification for the answer.', 'type': 'string'}
},
'required': ['answer']
}
}
llm = ChatOpenAI(model="gpt-3.5-turbo-0125", temperature=0)
structured_llm = llm.with_structured_output(oai_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.'
# }
Example: schema=Pydantic class, method="json_mode", include_raw=True:
.. code-block::
from langchain_openai import ChatOpenAI
@@ -1191,7 +1235,7 @@ class BaseChatOpenAI(BaseChatModel):
# 'parsing_error': None
# }
Example: JSON mode, no schema (schema=None, method="json_mode", include_raw=True):
Example: schema=None, method="json_mode", include_raw=True:
.. code-block::
structured_llm = llm.with_structured_output(method="json_mode", include_raw=True)
@@ -1209,8 +1253,6 @@ class BaseChatOpenAI(BaseChatModel):
# },
# 'parsing_error': None
# }
""" # noqa: E501
if kwargs:
raise ValueError(f"Received unsupported arguments {kwargs}")