github/langchain
1
0
Fork 0
mirror of https://github.com/hwchase17/langchain.git synced 2025-04-27 19:46:55 +00:00
Code Issues Packages Projects Releases Wiki Activity
141f8e5ba4
langchain/libs/partners/openai/langchain_openai/chat_models/base.py

2615 lines
103 KiB
Python
Raw Blame History

"""OpenAI chat wrapper."""
from __future__ import annotations
import base64
import json
import logging
import os
import re
import ssl
import sys
import warnings
from functools import partial
from io import BytesIO
from math import ceil
from operator import itemgetter
from typing import (
Any,
AsyncIterator,
Callable,
Dict,
Iterator,
List,
Literal,
Mapping,
Optional,
Sequence,
Tuple,
Type,
TypedDict,
TypeVar,
Union,
cast,
)
from urllib.parse import urlparse
import certifi
import openai
import tiktoken
from langchain_core._api.deprecation import deprecated
from langchain_core.callbacks import (
AsyncCallbackManagerForLLMRun,
CallbackManagerForLLMRun,
)
from langchain_core.language_models import LanguageModelInput
from langchain_core.language_models.chat_models import (
BaseChatModel,
LangSmithParams,
agenerate_from_stream,
generate_from_stream,
)
from langchain_core.messages import (
AIMessage,
AIMessageChunk,
BaseMessage,
BaseMessageChunk,
ChatMessage,
ChatMessageChunk,
FunctionMessage,
FunctionMessageChunk,
HumanMessage,
HumanMessageChunk,
InvalidToolCall,
SystemMessage,
SystemMessageChunk,
ToolCall,
ToolMessage,
ToolMessageChunk,
)
from langchain_core.messages.ai import (
InputTokenDetails,
OutputTokenDetails,
UsageMetadata,
)
from langchain_core.messages.tool import tool_call_chunk
from langchain_core.output_parsers import JsonOutputParser, PydanticOutputParser
from langchain_core.output_parsers.openai_tools import (
JsonOutputKeyToolsParser,
PydanticToolsParser,
make_invalid_tool_call,
parse_tool_call,
)
from langchain_core.outputs import ChatGeneration, ChatGenerationChunk, ChatResult
from langchain_core.runnables import (
Runnable,
RunnableLambda,
RunnableMap,
RunnablePassthrough,
)
from langchain_core.runnables.config import run_in_executor
from langchain_core.tools import BaseTool
from langchain_core.utils import get_pydantic_field_names
from langchain_core.utils.function_calling import (
convert_to_openai_function,
convert_to_openai_tool,
)
from langchain_core.utils.pydantic import (
PydanticBaseModel,
TypeBaseModel,
is_basemodel_subclass,
)
from langchain_core.utils.utils import _build_model_kwargs, from_env, secret_from_env
from pydantic import BaseModel, ConfigDict, Field, SecretStr, model_validator
from pydantic.v1 import BaseModel as BaseModelV1
from typing_extensions import Self
logger = logging.getLogger(__name__)
# This SSL context is equivelent to the default `verify=True`.
# https://www.python-httpx.org/advanced/ssl/#configuring-client-instances
global_ssl_context = ssl.create_default_context(cafile=certifi.where())
def _convert_dict_to_message(_dict: Mapping[str, Any]) -> BaseMessage:
"""Convert a dictionary to a LangChain message.
Args:
_dict: The dictionary.
Returns:
The LangChain message.
"""
role = _dict.get("role")
name = _dict.get("name")
id_ = _dict.get("id")
if role == "user":
return HumanMessage(content=_dict.get("content", ""), id=id_, name=name)
elif role == "assistant":
# Fix for azure
# Also OpenAI returns None for tool invocations
content = _dict.get("content", "") or ""
additional_kwargs: Dict = {}
if function_call := _dict.get("function_call"):
additional_kwargs["function_call"] = dict(function_call)
tool_calls = []
invalid_tool_calls = []
if raw_tool_calls := _dict.get("tool_calls"):
additional_kwargs["tool_calls"] = raw_tool_calls
for raw_tool_call in raw_tool_calls:
try:
tool_calls.append(parse_tool_call(raw_tool_call, return_id=True))
except Exception as e:
invalid_tool_calls.append(
make_invalid_tool_call(raw_tool_call, str(e))
)
if audio := _dict.get("audio"):
additional_kwargs["audio"] = audio
return AIMessage(
content=content,
additional_kwargs=additional_kwargs,
name=name,
id=id_,
tool_calls=tool_calls,
invalid_tool_calls=invalid_tool_calls,
)
elif role in ("system", "developer"):
if role == "developer":
additional_kwargs = {"__openai_role__": role}
else:
additional_kwargs = {}
return SystemMessage(
content=_dict.get("content", ""),
name=name,
id=id_,
additional_kwargs=additional_kwargs,
)
elif role == "function":
return FunctionMessage(
content=_dict.get("content", ""), name=cast(str, _dict.get("name")), id=id_
)
elif role == "tool":
additional_kwargs = {}
if "name" in _dict:
additional_kwargs["name"] = _dict["name"]
return ToolMessage(
content=_dict.get("content", ""),
tool_call_id=cast(str, _dict.get("tool_call_id")),
additional_kwargs=additional_kwargs,
name=name,
id=id_,
)
else:
return ChatMessage(content=_dict.get("content", ""), role=role, id=id_) # type: ignore[arg-type]
def _format_message_content(content: Any) -> Any:
"""Format message content."""
if content and isinstance(content, list):
# Remove unexpected block types
formatted_content = []
for block in content:
if (
isinstance(block, dict)
and "type" in block
and block["type"] == "tool_use"
):
continue
else:
formatted_content.append(block)
else:
formatted_content = content
return formatted_content
def _convert_message_to_dict(message: BaseMessage) -> dict:
"""Convert a LangChain message to a dictionary.
Args:
message: The LangChain message.
Returns:
The dictionary.
"""
message_dict: Dict[str, Any] = {"content": _format_message_content(message.content)}
if (name := message.name or message.additional_kwargs.get("name")) is not None:
message_dict["name"] = name
# populate role and additional message data
if isinstance(message, ChatMessage):
message_dict["role"] = message.role
elif isinstance(message, HumanMessage):
message_dict["role"] = "user"
elif isinstance(message, AIMessage):
message_dict["role"] = "assistant"
if "function_call" in message.additional_kwargs:
message_dict["function_call"] = message.additional_kwargs["function_call"]
if message.tool_calls or message.invalid_tool_calls:
message_dict["tool_calls"] = [
_lc_tool_call_to_openai_tool_call(tc) for tc in message.tool_calls
] + [
_lc_invalid_tool_call_to_openai_tool_call(tc)
for tc in message.invalid_tool_calls
]
elif "tool_calls" in message.additional_kwargs:
message_dict["tool_calls"] = message.additional_kwargs["tool_calls"]
tool_call_supported_props = {"id", "type", "function"}
message_dict["tool_calls"] = [
{k: v for k, v in tool_call.items() if k in tool_call_supported_props}
for tool_call in message_dict["tool_calls"]
]
else:
pass
# If tool calls present, content null value should be None not empty string.
if "function_call" in message_dict or "tool_calls" in message_dict:
message_dict["content"] = message_dict["content"] or None
if "audio" in message.additional_kwargs:
# openai doesn't support passing the data back - only the id
# https://platform.openai.com/docs/guides/audio/multi-turn-conversations
raw_audio = message.additional_kwargs["audio"]
audio = (
{"id": message.additional_kwargs["audio"]["id"]}
if "id" in raw_audio
else raw_audio
)
message_dict["audio"] = audio
elif isinstance(message, SystemMessage):
message_dict["role"] = message.additional_kwargs.get(
"__openai_role__", "system"
)
elif isinstance(message, FunctionMessage):
message_dict["role"] = "function"
elif isinstance(message, ToolMessage):
message_dict["role"] = "tool"
message_dict["tool_call_id"] = message.tool_call_id
supported_props = {"content", "role", "tool_call_id"}
message_dict = {k: v for k, v in message_dict.items() if k in supported_props}
else:
raise TypeError(f"Got unknown type {message}")
return message_dict
def _convert_delta_to_message_chunk(
_dict: Mapping[str, Any], default_class: Type[BaseMessageChunk]
) -> BaseMessageChunk:
id_ = _dict.get("id")
role = cast(str, _dict.get("role"))
content = cast(str, _dict.get("content") or "")
additional_kwargs: Dict = {}
if _dict.get("function_call"):
function_call = dict(_dict["function_call"])
if "name" in function_call and function_call["name"] is None:
function_call["name"] = ""
additional_kwargs["function_call"] = function_call
tool_call_chunks = []
if raw_tool_calls := _dict.get("tool_calls"):
additional_kwargs["tool_calls"] = raw_tool_calls
try:
tool_call_chunks = [
tool_call_chunk(
name=rtc["function"].get("name"),
args=rtc["function"].get("arguments"),
id=rtc.get("id"),
index=rtc["index"],
)
for rtc in raw_tool_calls
]
except KeyError:
pass
if role == "user" or default_class == HumanMessageChunk:
return HumanMessageChunk(content=content, id=id_)
elif role == "assistant" or default_class == AIMessageChunk:
return AIMessageChunk(
content=content,
additional_kwargs=additional_kwargs,
id=id_,
tool_call_chunks=tool_call_chunks, # type: ignore[arg-type]
)
elif role in ("system", "developer") or default_class == SystemMessageChunk:
if role == "developer":
additional_kwargs = {"__openai_role__": "developer"}
else:
additional_kwargs = {}
return SystemMessageChunk(
content=content, id=id_, additional_kwargs=additional_kwargs
)
elif role == "function" or default_class == FunctionMessageChunk:
return FunctionMessageChunk(content=content, name=_dict["name"], id=id_)
elif role == "tool" or default_class == ToolMessageChunk:
return ToolMessageChunk(
content=content, tool_call_id=_dict["tool_call_id"], id=id_
)
elif role or default_class == ChatMessageChunk:
return ChatMessageChunk(content=content, role=role, id=id_)
else:
return default_class(content=content, id=id_) # type: ignore
def _update_token_usage(
overall_token_usage: Union[int, dict], new_usage: Union[int, dict]
) -> Union[int, dict]:
# Token usage is either ints or dictionaries
# `reasoning_tokens` is nested inside `completion_tokens_details`
if isinstance(new_usage, int):
if not isinstance(overall_token_usage, int):
raise ValueError(
f"Got different types for token usage: "
f"{type(new_usage)} and {type(overall_token_usage)}"
)
return new_usage + overall_token_usage
elif isinstance(new_usage, dict):
if not isinstance(overall_token_usage, dict):
raise ValueError(
f"Got different types for token usage: "
f"{type(new_usage)} and {type(overall_token_usage)}"
)
return {
k: _update_token_usage(overall_token_usage.get(k, 0), v)
for k, v in new_usage.items()
}
else:
warnings.warn(f"Unexpected type for token usage: {type(new_usage)}")
return new_usage
def _handle_openai_bad_request(e: openai.BadRequestError) -> None:
if (
"'response_format' of type 'json_schema' is not supported with this model"
) in e.message:
message = (
"This model does not support OpenAI's structured output feature, which "
"is the default method for `with_structured_output` as of "
"langchain-openai==0.3. To use `with_structured_output` with this model, "
'specify `method="function_calling"`.'
)
warnings.warn(message)
raise e
elif "Invalid schema for response_format" in e.message:
message = (
"Invalid schema for OpenAI's structured output feature, which is the "
"default method for `with_structured_output` as of langchain-openai==0.3. "
'Specify `method="function_calling"` instead or update your schema. '
"See supported schemas: "
"https://platform.openai.com/docs/guides/structured-outputs#supported-schemas" # noqa: E501
)
warnings.warn(message)
raise e
else:
raise
class _FunctionCall(TypedDict):
name: str
_BM = TypeVar("_BM", bound=BaseModel)
_DictOrPydanticClass = Union[Dict[str, Any], Type[_BM], Type]
_DictOrPydantic = Union[Dict, _BM]
class _AllReturnType(TypedDict):
raw: BaseMessage
parsed: Optional[_DictOrPydantic]
parsing_error: Optional[BaseException]
class BaseChatOpenAI(BaseChatModel):
client: Any = Field(default=None, exclude=True) #: :meta private:
async_client: Any = Field(default=None, exclude=True) #: :meta private:
root_client: Any = Field(default=None, exclude=True) #: :meta private:
root_async_client: Any = Field(default=None, exclude=True) #: :meta private:
model_name: str = Field(default="gpt-3.5-turbo", alias="model")
"""Model name to use."""
temperature: Optional[float] = None
"""What sampling temperature to use."""
model_kwargs: Dict[str, Any] = Field(default_factory=dict)
"""Holds any model parameters valid for `create` call not explicitly specified."""
openai_api_key: Optional[SecretStr] = Field(
alias="api_key", default_factory=secret_from_env("OPENAI_API_KEY", default=None)
)
openai_api_base: Optional[str] = Field(default=None, alias="base_url")
"""Base URL path for API requests, leave blank if not using a proxy or service
emulator."""
openai_organization: Optional[str] = Field(default=None, alias="organization")
"""Automatically inferred from env var `OPENAI_ORG_ID` if not provided."""
# to support explicit proxy for OpenAI
openai_proxy: Optional[str] = Field(
default_factory=from_env("OPENAI_PROXY", default=None)
)
request_timeout: Union[float, Tuple[float, float], Any, None] = Field(
default=None, alias="timeout"
)
"""Timeout for requests to OpenAI completion API. Can be float, httpx.Timeout or
None."""
max_retries: Optional[int] = None
"""Maximum number of retries to make when generating."""
presence_penalty: Optional[float] = None
"""Penalizes repeated tokens."""
frequency_penalty: Optional[float] = None
"""Penalizes repeated tokens according to frequency."""
seed: Optional[int] = None
"""Seed for generation"""
logprobs: Optional[bool] = None
"""Whether to return logprobs."""
top_logprobs: Optional[int] = None
"""Number of most likely tokens to return at each token position, each with
an associated log probability. `logprobs` must be set to true
if this parameter is used."""
logit_bias: Optional[Dict[int, int]] = None
"""Modify the likelihood of specified tokens appearing in the completion."""
streaming: bool = False
"""Whether to stream the results or not."""
n: Optional[int] = None
"""Number of chat completions to generate for each prompt."""
top_p: Optional[float] = None
"""Total probability mass of tokens to consider at each step."""
max_tokens: Optional[int] = Field(default=None)
"""Maximum number of tokens to generate."""
reasoning_effort: Optional[str] = None
"""Constrains effort on reasoning for reasoning models.
Reasoning models only, like OpenAI o1 and o3-mini.
Currently supported values are low, medium, and high. Reducing reasoning effort
can result in faster responses and fewer tokens used on reasoning in a response.
.. versionadded:: 0.2.14
"""
tiktoken_model_name: Optional[str] = None
"""The model name to pass to tiktoken when using this class.
Tiktoken is used to count the number of tokens in documents to constrain
them to be under a certain limit. By default, when set to None, this will
be the same as the embedding model name. However, there are some cases
where you may want to use this Embedding class with a model name not
supported by tiktoken. This can include when using Azure embeddings or
when using one of the many model providers that expose an OpenAI-like
API but with different models. In those cases, in order to avoid erroring
when tiktoken is called, you can specify a model name to use here."""
default_headers: Union[Mapping[str, str], None] = None
default_query: Union[Mapping[str, object], None] = None
# Configure a custom httpx client. See the
# [httpx documentation](https://www.python-httpx.org/api/#client) for more details.
http_client: Union[Any, None] = Field(default=None, exclude=True)
"""Optional httpx.Client. Only used for sync invocations. Must specify
http_async_client as well if you'd like a custom client for async invocations.
"""
http_async_client: Union[Any, None] = Field(default=None, exclude=True)
"""Optional httpx.AsyncClient. Only used for async invocations. Must specify
http_client as well if you'd like a custom client for sync invocations."""
stop: Optional[Union[List[str], str]] = Field(default=None, alias="stop_sequences")
"""Default stop sequences."""
extra_body: Optional[Mapping[str, Any]] = None
"""Optional additional JSON properties to include in the request parameters when
making requests to OpenAI compatible APIs, such as vLLM."""
include_response_headers: bool = False
"""Whether to include response headers in the output message response_metadata."""
disabled_params: Optional[Dict[str, Any]] = Field(default=None)
"""Parameters of the OpenAI client or chat.completions endpoint that should be
disabled for the given model.
Should be specified as ``{"param": None | ['val1', 'val2']}`` where the key is the
parameter and the value is either None, meaning that parameter should never be
used, or it's a list of disabled values for the parameter.
For example, older models may not support the 'parallel_tool_calls' parameter at
all, in which case ``disabled_params={"parallel_tool_calls": None}`` can ben passed
in.
If a parameter is disabled then it will not be used by default in any methods, e.g.
in :meth:`~langchain_openai.chat_models.base.ChatOpenAI.with_structured_output`.
However this does not prevent a user from directly passed in the parameter during
invocation.
"""
model_config = ConfigDict(populate_by_name=True)
@model_validator(mode="before")
@classmethod
def build_extra(cls, values: Dict[str, Any]) -> Any:
"""Build extra kwargs from additional params that were passed in."""
all_required_field_names = get_pydantic_field_names(cls)
values = _build_model_kwargs(values, all_required_field_names)
return values
@model_validator(mode="before")
@classmethod
def validate_temperature(cls, values: Dict[str, Any]) -> Any:
"""Currently o1 models only allow temperature=1."""
model = values.get("model_name") or values.get("model") or ""
if model.startswith("o1") and "temperature" not in values:
values["temperature"] = 1
return values
@model_validator(mode="after")
def validate_environment(self) -> Self:
"""Validate that api key and python package exists in environment."""
if self.n is not None and self.n < 1:
raise ValueError("n must be at least 1.")
elif self.n is not None and self.n > 1 and self.streaming:
raise ValueError("n must be 1 when streaming.")
# Check OPENAI_ORGANIZATION for backwards compatibility.
self.openai_organization = (
self.openai_organization
or os.getenv("OPENAI_ORG_ID")
or os.getenv("OPENAI_ORGANIZATION")
)
self.openai_api_base = self.openai_api_base or os.getenv("OPENAI_API_BASE")
client_params: dict = {
"api_key": (
self.openai_api_key.get_secret_value() if self.openai_api_key else None
),
"organization": self.openai_organization,
"base_url": self.openai_api_base,
"timeout": self.request_timeout,
"default_headers": self.default_headers,
"default_query": self.default_query,
}
if self.max_retries is not None:
client_params["max_retries"] = self.max_retries
if self.openai_proxy and (self.http_client or self.http_async_client):
openai_proxy = self.openai_proxy
http_client = self.http_client
http_async_client = self.http_async_client
raise ValueError(
"Cannot specify 'openai_proxy' if one of "
"'http_client'/'http_async_client' is already specified. Received:\n"
f"{openai_proxy=}\n{http_client=}\n{http_async_client=}"
)
if not self.client:
if self.openai_proxy and not self.http_client:
try:
import httpx
except ImportError as e:
raise ImportError(
"Could not import httpx python package. "
"Please install it with `pip install httpx`."
) from e
self.http_client = httpx.Client(
proxy=self.openai_proxy, verify=global_ssl_context
)
sync_specific = {"http_client": self.http_client}
self.root_client = openai.OpenAI(**client_params, **sync_specific) # type: ignore[arg-type]
self.client = self.root_client.chat.completions
if not self.async_client:
if self.openai_proxy and not self.http_async_client:
try:
import httpx
except ImportError as e:
raise ImportError(
"Could not import httpx python package. "
"Please install it with `pip install httpx`."
) from e
self.http_async_client = httpx.AsyncClient(
proxy=self.openai_proxy, verify=global_ssl_context
)
async_specific = {"http_client": self.http_async_client}
self.root_async_client = openai.AsyncOpenAI(
**client_params,
**async_specific, # type: ignore[arg-type]
)
self.async_client = self.root_async_client.chat.completions
return self
@property
def _default_params(self) -> Dict[str, Any]:
"""Get the default parameters for calling OpenAI API."""
exclude_if_none = {
"presence_penalty": self.presence_penalty,
"frequency_penalty": self.frequency_penalty,
"seed": self.seed,
"top_p": self.top_p,
"logprobs": self.logprobs,
"top_logprobs": self.top_logprobs,
"logit_bias": self.logit_bias,
"stop": self.stop or None, # also exclude empty list for this
"max_tokens": self.max_tokens,
"extra_body": self.extra_body,
"n": self.n,
"temperature": self.temperature,
"reasoning_effort": self.reasoning_effort,
}
params = {
"model": self.model_name,
"stream": self.streaming,
**{k: v for k, v in exclude_if_none.items() if v is not None},
**self.model_kwargs,
}
return params
def _combine_llm_outputs(self, llm_outputs: List[Optional[dict]]) -> dict:
overall_token_usage: dict = {}
system_fingerprint = None
for output in llm_outputs:
if output is None:
# Happens in streaming
continue
token_usage = output["token_usage"]
if token_usage is not None:
for k, v in token_usage.items():
if v is None:
continue
if k in overall_token_usage:
overall_token_usage[k] = _update_token_usage(
overall_token_usage[k], v
)
else:
overall_token_usage[k] = v
if system_fingerprint is None:
system_fingerprint = output.get("system_fingerprint")
combined = {"token_usage": overall_token_usage, "model_name": self.model_name}
if system_fingerprint:
combined["system_fingerprint"] = system_fingerprint
return combined
def _convert_chunk_to_generation_chunk(
self,
chunk: dict,
default_chunk_class: Type,
base_generation_info: Optional[Dict],
) -> Optional[ChatGenerationChunk]:
if chunk.get("type") == "content.delta": # from beta.chat.completions.stream
return None
token_usage = chunk.get("usage")
choices = (
chunk.get("choices", [])
# from beta.chat.completions.stream
or chunk.get("chunk", {}).get("choices", [])
)
usage_metadata: Optional[UsageMetadata] = (
_create_usage_metadata(token_usage) if token_usage else None
)
if len(choices) == 0:
# logprobs is implicitly None
generation_chunk = ChatGenerationChunk(
message=default_chunk_class(content="", usage_metadata=usage_metadata)
)
return generation_chunk
choice = choices[0]
if choice["delta"] is None:
return None
message_chunk = _convert_delta_to_message_chunk(
choice["delta"], default_chunk_class
)
generation_info = {**base_generation_info} if base_generation_info else {}
if finish_reason := choice.get("finish_reason"):
generation_info["finish_reason"] = finish_reason
if model_name := chunk.get("model"):
generation_info["model_name"] = model_name
if system_fingerprint := chunk.get("system_fingerprint"):
generation_info["system_fingerprint"] = system_fingerprint
logprobs = choice.get("logprobs")
if logprobs:
generation_info["logprobs"] = logprobs
if usage_metadata and isinstance(message_chunk, AIMessageChunk):
message_chunk.usage_metadata = usage_metadata
generation_chunk = ChatGenerationChunk(
message=message_chunk, generation_info=generation_info or None
)
return generation_chunk
def _stream(
self,
messages: List[BaseMessage],
stop: Optional[List[str]] = None,
run_manager: Optional[CallbackManagerForLLMRun] = None,
**kwargs: Any,
) -> Iterator[ChatGenerationChunk]:
kwargs["stream"] = True
payload = self._get_request_payload(messages, stop=stop, **kwargs)
default_chunk_class: Type[BaseMessageChunk] = AIMessageChunk
base_generation_info = {}
if "response_format" in payload:
if self.include_response_headers:
warnings.warn(
"Cannot currently include response headers when response_format is "
"specified."
)
payload.pop("stream")
response_stream = self.root_client.beta.chat.completions.stream(**payload)
context_manager = response_stream
else:
if self.include_response_headers:
raw_response = self.client.with_raw_response.create(**payload)
response = raw_response.parse()
base_generation_info = {"headers": dict(raw_response.headers)}
else:
response = self.client.create(**payload)
context_manager = response
try:
with context_manager as response:
is_first_chunk = True
for chunk in response:
if not isinstance(chunk, dict):
chunk = chunk.model_dump()
generation_chunk = self._convert_chunk_to_generation_chunk(
chunk,
default_chunk_class,
base_generation_info if is_first_chunk else {},
)
if generation_chunk is None:
continue
default_chunk_class = generation_chunk.message.__class__
logprobs = (generation_chunk.generation_info or {}).get("logprobs")
if run_manager:
run_manager.on_llm_new_token(
generation_chunk.text,
chunk=generation_chunk,
logprobs=logprobs,
)
is_first_chunk = False
yield generation_chunk
except openai.BadRequestError as e:
_handle_openai_bad_request(e)
if hasattr(response, "get_final_completion") and "response_format" in payload:
final_completion = response.get_final_completion()
generation_chunk = self._get_generation_chunk_from_completion(
final_completion
)
if run_manager:
run_manager.on_llm_new_token(
generation_chunk.text, chunk=generation_chunk
)
yield generation_chunk
def _generate(
self,
messages: List[BaseMessage],
stop: Optional[List[str]] = None,
run_manager: Optional[CallbackManagerForLLMRun] = None,
**kwargs: Any,
) -> ChatResult:
if self.streaming:
stream_iter = self._stream(
messages, stop=stop, run_manager=run_manager, **kwargs
)
return generate_from_stream(stream_iter)
payload = self._get_request_payload(messages, stop=stop, **kwargs)
generation_info = None
if "response_format" in payload:
if self.include_response_headers:
warnings.warn(
"Cannot currently include response headers when response_format is "
"specified."
)
payload.pop("stream")
try:
response = self.root_client.beta.chat.completions.parse(**payload)
except openai.BadRequestError as e:
_handle_openai_bad_request(e)
elif self.include_response_headers:
raw_response = self.client.with_raw_response.create(**payload)
response = raw_response.parse()
generation_info = {"headers": dict(raw_response.headers)}
else:
response = self.client.create(**payload)
return self._create_chat_result(response, generation_info)
def _get_request_payload(
self,
input_: LanguageModelInput,
*,
stop: Optional[List[str]] = None,
**kwargs: Any,
) -> dict:
messages = self._convert_input(input_).to_messages()
if stop is not None:
kwargs["stop"] = stop
return {
"messages": [_convert_message_to_dict(m) for m in messages],
**self._default_params,
**kwargs,
}
def _create_chat_result(
self,
response: Union[dict, openai.BaseModel],
generation_info: Optional[Dict] = None,
) -> ChatResult:
generations = []
response_dict = (
response if isinstance(response, dict) else response.model_dump()
)
# Sometimes the AI Model calling will get error, we should raise it.
# Otherwise, the next code 'choices.extend(response["choices"])'
# will throw a "TypeError: 'NoneType' object is not iterable" error
# to mask the true error. Because 'response["choices"]' is None.
if response_dict.get("error"):
raise ValueError(response_dict.get("error"))
token_usage = response_dict.get("usage")
for res in response_dict["choices"]:
message = _convert_dict_to_message(res["message"])
if token_usage and isinstance(message, AIMessage):
message.usage_metadata = _create_usage_metadata(token_usage)
generation_info = generation_info or {}
generation_info["finish_reason"] = (
res.get("finish_reason")
if res.get("finish_reason") is not None
else generation_info.get("finish_reason")
)
if "logprobs" in res:
generation_info["logprobs"] = res["logprobs"]
gen = ChatGeneration(message=message, generation_info=generation_info)
generations.append(gen)
llm_output = {
"token_usage": token_usage,
"model_name": response_dict.get("model", self.model_name),
"system_fingerprint": response_dict.get("system_fingerprint", ""),
}
if isinstance(response, openai.BaseModel) and getattr(
response, "choices", None
):
message = response.choices[0].message # type: ignore[attr-defined]
if hasattr(message, "parsed"):
generations[0].message.additional_kwargs["parsed"] = message.parsed
if hasattr(message, "refusal"):
generations[0].message.additional_kwargs["refusal"] = message.refusal
return ChatResult(generations=generations, llm_output=llm_output)
async def _astream(
self,
messages: List[BaseMessage],
stop: Optional[List[str]] = None,
run_manager: Optional[AsyncCallbackManagerForLLMRun] = None,
**kwargs: Any,
) -> AsyncIterator[ChatGenerationChunk]:
kwargs["stream"] = True
payload = self._get_request_payload(messages, stop=stop, **kwargs)
default_chunk_class: Type[BaseMessageChunk] = AIMessageChunk
base_generation_info = {}
if "response_format" in payload:
if self.include_response_headers:
warnings.warn(
"Cannot currently include response headers when response_format is "
"specified."
)
payload.pop("stream")
response_stream = self.root_async_client.beta.chat.completions.stream(
**payload
)
context_manager = response_stream
else:
if self.include_response_headers:
raw_response = await self.async_client.with_raw_response.create(
**payload
)
response = raw_response.parse()
base_generation_info = {"headers": dict(raw_response.headers)}
else:
response = await self.async_client.create(**payload)
context_manager = response
try:
async with context_manager as response:
is_first_chunk = True
async for chunk in response:
if not isinstance(chunk, dict):
chunk = chunk.model_dump()
generation_chunk = self._convert_chunk_to_generation_chunk(
chunk,
default_chunk_class,
base_generation_info if is_first_chunk else {},
)
if generation_chunk is None:
continue
default_chunk_class = generation_chunk.message.__class__
logprobs = (generation_chunk.generation_info or {}).get("logprobs")
if run_manager:
await run_manager.on_llm_new_token(
generation_chunk.text,
chunk=generation_chunk,
logprobs=logprobs,
)
is_first_chunk = False
yield generation_chunk
except openai.BadRequestError as e:
_handle_openai_bad_request(e)
if hasattr(response, "get_final_completion") and "response_format" in payload:
final_completion = await response.get_final_completion()
generation_chunk = self._get_generation_chunk_from_completion(
final_completion
)
if run_manager:
await run_manager.on_llm_new_token(
generation_chunk.text, chunk=generation_chunk
)
yield generation_chunk
async def _agenerate(
self,
messages: List[BaseMessage],
stop: Optional[List[str]] = None,
run_manager: Optional[AsyncCallbackManagerForLLMRun] = None,
**kwargs: Any,
) -> ChatResult:
if self.streaming:
stream_iter = self._astream(
messages, stop=stop, run_manager=run_manager, **kwargs
)
return await agenerate_from_stream(stream_iter)
payload = self._get_request_payload(messages, stop=stop, **kwargs)
generation_info = None
if "response_format" in payload:
if self.include_response_headers:
warnings.warn(
"Cannot currently include response headers when response_format is "
"specified."
)
payload.pop("stream")
try:
response = await self.root_async_client.beta.chat.completions.parse(
**payload
)
except openai.BadRequestError as e:
_handle_openai_bad_request(e)
elif self.include_response_headers:
raw_response = await self.async_client.with_raw_response.create(**payload)
response = raw_response.parse()
generation_info = {"headers": dict(raw_response.headers)}
else:
response = await self.async_client.create(**payload)
return await run_in_executor(
None, self._create_chat_result, response, generation_info
)
@property
def _identifying_params(self) -> Dict[str, Any]:
"""Get the identifying parameters."""
return {"model_name": self.model_name, **self._default_params}
def _get_invocation_params(
self, stop: Optional[List[str]] = None, **kwargs: Any
) -> Dict[str, Any]:
"""Get the parameters used to invoke the model."""
return {
"model": self.model_name,
**super()._get_invocation_params(stop=stop),
**self._default_params,
**kwargs,
}
def _get_ls_params(
self, stop: Optional[List[str]] = None, **kwargs: Any
) -> LangSmithParams:
"""Get standard params for tracing."""
params = self._get_invocation_params(stop=stop, **kwargs)
ls_params = LangSmithParams(
ls_provider="openai",
ls_model_name=self.model_name,
ls_model_type="chat",
ls_temperature=params.get("temperature", self.temperature),
)
if ls_max_tokens := params.get("max_tokens", self.max_tokens) or params.get(
"max_completion_tokens", self.max_tokens
):
ls_params["ls_max_tokens"] = ls_max_tokens
if ls_stop := stop or params.get("stop", None):
ls_params["ls_stop"] = ls_stop
return ls_params
@property
def _llm_type(self) -> str:
"""Return type of chat model."""
return "openai-chat"
def _get_encoding_model(self) -> Tuple[str, tiktoken.Encoding]:
if self.tiktoken_model_name is not None:
model = self.tiktoken_model_name
else:
model = self.model_name
try:
encoding = tiktoken.encoding_for_model(model)
except KeyError:
model = "cl100k_base"
encoding = tiktoken.get_encoding(model)
return model, encoding
def get_token_ids(self, text: str) -> List[int]:
"""Get the tokens present in the text with tiktoken package."""
if self.custom_get_token_ids is not None:
return self.custom_get_token_ids(text)
# tiktoken NOT supported for Python 3.7 or below
if sys.version_info[1] <= 7:
return super().get_token_ids(text)
_, encoding_model = self._get_encoding_model()
return encoding_model.encode(text)
def get_num_tokens_from_messages(
self,
messages: List[BaseMessage],
tools: Optional[
Sequence[Union[Dict[str, Any], Type, Callable, BaseTool]]
] = None,
) -> int:
"""Calculate num tokens for gpt-3.5-turbo and gpt-4 with tiktoken package.
**Requirements**: You must have the ``pillow`` installed if you want to count
image tokens if you are specifying the image as a base64 string, and you must
have both ``pillow`` and ``httpx`` installed if you are specifying the image
as a URL. If these aren't installed image inputs will be ignored in token
counting.
OpenAI reference: https://github.com/openai/openai-cookbook/blob/
main/examples/How_to_format_inputs_to_ChatGPT_models.ipynb
Args:
messages: The message inputs to tokenize.
tools: If provided, sequence of dict, BaseModel, function, or BaseTools
to be converted to tool schemas.
"""
# TODO: Count bound tools as part of input.
if tools is not None:
warnings.warn(
"Counting tokens in tool schemas is not yet supported. Ignoring tools."
)
if sys.version_info[1] <= 7:
return super().get_num_tokens_from_messages(messages)
model, encoding = self._get_encoding_model()
if model.startswith("gpt-3.5-turbo-0301"):
# every message follows <im_start>{role/name}\n{content}<im_end>\n
tokens_per_message = 4
# if there's a name, the role is omitted
tokens_per_name = -1
elif model.startswith("gpt-3.5-turbo") or model.startswith("gpt-4"):
tokens_per_message = 3
tokens_per_name = 1
else:
raise NotImplementedError(
f"get_num_tokens_from_messages() is not presently implemented "
f"for model {model}. See "
"https://platform.openai.com/docs/guides/text-generation/managing-tokens" # noqa: E501
" for information on how messages are converted to tokens."
)
num_tokens = 0
messages_dict = [_convert_message_to_dict(m) for m in messages]
for message in messages_dict:
num_tokens += tokens_per_message
for key, value in message.items():
# This is an inferred approximation. OpenAI does not document how to
# count tool message tokens.
if key == "tool_call_id":
num_tokens += 3
continue
if isinstance(value, list):
# content or tool calls
for val in value:
if isinstance(val, str) or val["type"] == "text":
text = val["text"] if isinstance(val, dict) else val
num_tokens += len(encoding.encode(text))
elif val["type"] == "image_url":
if val["image_url"].get("detail") == "low":
num_tokens += 85
else:
image_size = _url_to_size(val["image_url"]["url"])
if not image_size:
continue
num_tokens += _count_image_tokens(*image_size)
# Tool/function call token counting is not documented by OpenAI.
# This is an approximation.
elif val["type"] == "function":
num_tokens += len(
encoding.encode(val["function"]["arguments"])
)
num_tokens += len(encoding.encode(val["function"]["name"]))
else:
raise ValueError(
f"Unrecognized content block type\n\n{val}"
)
elif not value:
continue
else:
# Cast str(value) in case the message value is not a string
# This occurs with function messages
num_tokens += len(encoding.encode(str(value)))
if key == "name":
num_tokens += tokens_per_name
# every reply is primed with <im_start>assistant
num_tokens += 3
return num_tokens
@deprecated(
since="0.2.1",
alternative="langchain_openai.chat_models.base.ChatOpenAI.bind_tools",
removal="1.0.0",
)
def bind_functions(
self,
functions: Sequence[Union[Dict[str, Any], Type[BaseModel], Callable, BaseTool]],
function_call: Optional[
Union[_FunctionCall, str, Literal["auto", "none"]]
] = None,
**kwargs: Any,
) -> Runnable[LanguageModelInput, BaseMessage]:
"""Bind functions (and other objects) to this chat model.
Assumes model is compatible with OpenAI function-calling API.
NOTE: Using bind_tools is recommended instead, as the `functions` and
`function_call` request parameters are officially marked as deprecated by
OpenAI.
Args:
functions: A list of function definitions to bind to this chat model.
Can be a dictionary, pydantic model, or callable. Pydantic
models and callables will be automatically converted to
their schema dictionary representation.
function_call: Which function to require the model to call.
Must be the name of the single provided function or
"auto" to automatically determine which function to call
(if any).
**kwargs: Any additional parameters to pass to the
:class:`~langchain.runnable.Runnable` constructor.
"""
formatted_functions = [convert_to_openai_function(fn) for fn in functions]
if function_call is not None:
function_call = (
{"name": function_call}
if isinstance(function_call, str)
and function_call not in ("auto", "none")
else function_call
)
if isinstance(function_call, dict) and len(formatted_functions) != 1:
raise ValueError(
"When specifying `function_call`, you must provide exactly one "
"function."
)
if (
isinstance(function_call, dict)
and formatted_functions[0]["name"] != function_call["name"]
):
raise ValueError(
f"Function call {function_call} was specified, but the only "
f"provided function was {formatted_functions[0]['name']}."
)
kwargs = {**kwargs, "function_call": function_call}
return super().bind(functions=formatted_functions, **kwargs)
def bind_tools(
self,
tools: Sequence[Union[Dict[str, Any], Type, Callable, BaseTool]],
*,
tool_choice: Optional[
Union[dict, str, Literal["auto", "none", "required", "any"], bool]
] = None,
strict: Optional[bool] = None,
parallel_tool_calls: Optional[bool] = None,
**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.
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:
- 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.
strict: If True, model output is guaranteed to exactly match the JSON Schema
provided in the tool definition. If True, the input schema will be
validated according to
https://platform.openai.com/docs/guides/structured-outputs/supported-schemas.
If False, input schema will not be validated and model output will not
be validated.
If None, ``strict`` argument will not be passed to the model.
parallel_tool_calls: Set to ``False`` to disable parallel tool use.
Defaults to ``None`` (no specification, which allows parallel tool use).
kwargs: Any additional parameters are passed directly to
:meth:`~langchain_openai.chat_models.base.ChatOpenAI.bind`.
.. versionchanged:: 0.1.21
Support for ``strict`` argument added.
""" # noqa: E501
if parallel_tool_calls is not None:
kwargs["parallel_tool_calls"] = parallel_tool_calls
formatted_tools = [
convert_to_openai_tool(tool, strict=strict) for tool in tools
]
if tool_choice:
if isinstance(tool_choice, str):
# tool_choice is a tool/function name
if tool_choice not in ("auto", "none", "any", "required"):
tool_choice = {
"type": "function",
"function": {"name": tool_choice},
}
# 'any' is not natively supported by OpenAI API.
# We support 'any' since other models use this instead of 'required'.
if tool_choice == "any":
tool_choice = "required"
elif isinstance(tool_choice, bool):
tool_choice = "required"
elif isinstance(tool_choice, dict):
tool_names = [
formatted_tool["function"]["name"]
for formatted_tool in formatted_tools
]
if not any(
tool_name == tool_choice["function"]["name"]
for tool_name in tool_names
):
raise ValueError(
f"Tool choice {tool_choice} was specified, but the only "
f"provided tools were {tool_names}."
)
else:
raise ValueError(
f"Unrecognized tool_choice type. Expected str, bool or dict. "
f"Received: {tool_choice}"
)
kwargs["tool_choice"] = tool_choice
return super().bind(tools=formatted_tools, **kwargs)
def with_structured_output(
self,
schema: Optional[_DictOrPydanticClass] = None,
*,
method: Literal[
"function_calling", "json_mode", "json_schema"
] = "function_calling",
include_raw: bool = False,
strict: Optional[bool] = None,
**kwargs: Any,
) -> Runnable[LanguageModelInput, _DictOrPydantic]:
"""Model wrapper that returns outputs formatted to match the given schema.
Args:
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.
method: The method for steering model generation, one of:
- "function_calling":
Uses OpenAI's tool-calling (formerly called function calling)
API: https://platform.openai.com/docs/guides/function-calling
- "json_schema":
Uses OpenAI's Structured Output API: https://platform.openai.com/docs/guides/structured-outputs
Supported for "gpt-4o-mini", "gpt-4o-2024-08-06", "o1", and later
models.
- "json_mode":
Uses OpenAI's JSON mode. Note that if using JSON mode then you
must include instructions for formatting the output into the
desired schema into the model call:
https://platform.openai.com/docs/guides/structured-outputs/json-mode
Learn more about the differences between the methods and which models
support which methods here:
- https://platform.openai.com/docs/guides/structured-outputs/structured-outputs-vs-json-mode
- https://platform.openai.com/docs/guides/structured-outputs/function-calling-vs-response-format
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".
strict:
- True:
Model output is guaranteed to exactly match the schema.
The input schema will also be validated according to
https://platform.openai.com/docs/guides/structured-outputs/supported-schemas
- False:
Input schema will not be validated and model output will not be
validated.
- None:
``strict`` argument will not be passed to the model.
kwargs: Additional keyword args aren't supported.
Returns:
A Runnable that takes same inputs as a :class:`langchain_core.language_models.chat.BaseChatModel`.
| If ``include_raw`` is False and ``schema`` is a Pydantic class, Runnable outputs an instance of ``schema`` (i.e., a Pydantic object). Otherwise, if ``include_raw`` is False then Runnable outputs a dict.
| 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]
.. versionchanged:: 0.1.20
Added support for TypedDict class ``schema``.
.. versionchanged:: 0.1.21
Support for ``strict`` argument added.
Support for ``method`` = "json_schema" added.
""" # noqa: E501
if kwargs:
raise ValueError(f"Received unsupported arguments {kwargs}")
if strict is not None and method == "json_mode":
raise ValueError(
"Argument `strict` is not supported with `method`='json_mode'"
)
is_pydantic_schema = _is_pydantic_class(schema)
if method == "json_schema":
# Check for Pydantic BaseModel V1
if (
is_pydantic_schema and issubclass(schema, BaseModelV1) # type: ignore[arg-type]
):
warnings.warn(
"Received a Pydantic BaseModel V1 schema. This is not supported by "
'method="json_schema". Please use method="function_calling" '
"or specify schema via JSON Schema or Pydantic V2 BaseModel. "
'Overriding to method="function_calling".'
)
method = "function_calling"
# Check for incompatible model
if self.model_name and (
self.model_name.startswith("gpt-3")
or self.model_name.startswith("gpt-4-")
or self.model_name == "gpt-4"
):
warnings.warn(
f"Cannot use method='json_schema' with model {self.model_name} "
f"since it doesn't support OpenAI's Structured Output API. You can "
f"see supported models here: "
f"https://platform.openai.com/docs/guides/structured-outputs#supported-models. " # noqa: E501
"To fix this warning, set `method='function_calling'. "
"Overriding to method='function_calling'."
)
method = "function_calling"
if method == "function_calling":
if schema is None:
raise ValueError(
"schema must be specified when method is not 'json_mode'. "
"Received None."
)
tool_name = convert_to_openai_tool(schema)["function"]["name"]
bind_kwargs = self._filter_disabled_params(
tool_choice=tool_name,
parallel_tool_calls=False,
strict=strict,
structured_output_format={
"kwargs": {"method": method},
"schema": _convert_to_json_schema(schema, strict=strict),
},
)
llm = self.bind_tools([schema], **bind_kwargs)
if is_pydantic_schema:
output_parser: Runnable = PydanticToolsParser(
tools=[schema], # type: ignore[list-item]
first_tool_only=True, # type: ignore[list-item]
)
else:
output_parser = JsonOutputKeyToolsParser(
key_name=tool_name, first_tool_only=True
)
elif method == "json_mode":
llm = self.bind(
response_format={"type": "json_object"},
structured_output_format={
"kwargs": {"method": method},
"schema": _convert_to_json_schema(schema, strict=strict)
if schema
else None,
},
)
output_parser = (
PydanticOutputParser(pydantic_object=schema) # type: ignore[arg-type]
if is_pydantic_schema
else JsonOutputParser()
)
elif method == "json_schema":
if schema is None:
raise ValueError(
"schema must be specified when method is not 'json_mode'. "
"Received None."
)
response_format = _convert_to_openai_response_format(schema, strict=strict)
llm = self.bind(
response_format=response_format,
structured_output_format={
"kwargs": {"method": method},
"schema": _convert_to_json_schema(schema, strict=strict),
},
)
if is_pydantic_schema:
output_parser = RunnableLambda(
partial(_oai_structured_outputs_parser, schema=cast(type, schema))
).with_types(output_type=cast(type, schema))
else:
output_parser = JsonOutputParser()
else:
raise ValueError(
f"Unrecognized method argument. Expected one of 'function_calling' or "
f"'json_mode'. Received: '{method}'"
)
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
def _filter_disabled_params(self, **kwargs: Any) -> Dict[str, Any]:
if not self.disabled_params:
return kwargs
filtered = {}
for k, v in kwargs.items():
# Skip param
if k in self.disabled_params and (
self.disabled_params[k] is None or v in self.disabled_params[k]
):
continue
# Keep param
else:
filtered[k] = v
return filtered
def _get_generation_chunk_from_completion(
self, completion: openai.BaseModel
) -> ChatGenerationChunk:
"""Get chunk from completion (e.g., from final completion of a stream)."""
chat_result = self._create_chat_result(completion)
chat_message = chat_result.generations[0].message
if isinstance(chat_message, AIMessage):
usage_metadata = chat_message.usage_metadata
# Skip tool_calls, already sent as chunks
if "tool_calls" in chat_message.additional_kwargs:
chat_message.additional_kwargs.pop("tool_calls")
else:
usage_metadata = None
message = AIMessageChunk(
content="",
additional_kwargs=chat_message.additional_kwargs,
usage_metadata=usage_metadata,
)
return ChatGenerationChunk(
message=message, generation_info=chat_result.llm_output
)
class ChatOpenAI(BaseChatOpenAI): # type: ignore[override]
"""OpenAI chat model integration.
.. dropdown:: Setup
:open:
Install ``langchain-openai`` and set environment variable ``OPENAI_API_KEY``.
.. code-block:: bash
pip install -U langchain-openai
export OPENAI_API_KEY="your-api-key"
.. dropdown:: Key init args — completion params
model: str
Name of OpenAI model to use.
temperature: float
Sampling temperature.
max_tokens: Optional[int]
Max number of tokens to generate.
logprobs: Optional[bool]
Whether to return logprobs.
stream_options: Dict
Configure streaming outputs, like whether to return token usage when
streaming (``{"include_usage": True}``).
See full list of supported init args and their descriptions in the params section.
.. dropdown:: Key init args — client params
timeout: Union[float, Tuple[float, float], Any, None]
Timeout for requests.
max_retries: Optional[int]
Max number of retries.
api_key: Optional[str]
OpenAI API key. If not passed in will be read from env var OPENAI_API_KEY.
base_url: Optional[str]
Base URL for API requests. Only specify if using a proxy or service
emulator.
organization: Optional[str]
OpenAI organization ID. If not passed in will be read from env
var OPENAI_ORG_ID.
See full list of supported init args and their descriptions in the params section.
.. dropdown:: Instantiate
.. code-block:: python
from langchain_openai import ChatOpenAI
llm = ChatOpenAI(
model="gpt-4o",
temperature=0,
max_tokens=None,
timeout=None,
max_retries=2,
# api_key="...",
# base_url="...",
# organization="...",
# other params...
)
**NOTE**: Any param which is not explicitly supported will be passed directly to the
``openai.OpenAI.chat.completions.create(...)`` API every time to the model is
invoked. For example:
.. code-block:: python
from langchain_openai import ChatOpenAI
import openai
ChatOpenAI(..., frequency_penalty=0.2).invoke(...)
# results in underlying API call of:
openai.OpenAI(..).chat.completions.create(..., frequency_penalty=0.2)
# which is also equivalent to:
ChatOpenAI(...).invoke(..., frequency_penalty=0.2)
.. dropdown:: Invoke
.. code-block:: python
messages = [
(
"system",
"You are a helpful translator. Translate the user sentence to French.",
),
("human", "I love programming."),
]
llm.invoke(messages)
.. code-block:: pycon
AIMessage(
content="J'adore la programmation.",
response_metadata={
"token_usage": {
"completion_tokens": 5,
"prompt_tokens": 31,
"total_tokens": 36,
},
"model_name": "gpt-4o",
"system_fingerprint": "fp_43dfabdef1",
"finish_reason": "stop",
"logprobs": None,
},
id="run-012cffe2-5d3d-424d-83b5-51c6d4a593d1-0",
usage_metadata={"input_tokens": 31, "output_tokens": 5, "total_tokens": 36},
)
.. dropdown:: Stream
.. code-block:: python
for chunk in llm.stream(messages):
print(chunk.text(), end="")
.. code-block:: python
AIMessageChunk(content="", id="run-9e1517e3-12bf-48f2-bb1b-2e824f7cd7b0")
AIMessageChunk(content="J", id="run-9e1517e3-12bf-48f2-bb1b-2e824f7cd7b0")
AIMessageChunk(content="'adore", id="run-9e1517e3-12bf-48f2-bb1b-2e824f7cd7b0")
AIMessageChunk(content=" la", id="run-9e1517e3-12bf-48f2-bb1b-2e824f7cd7b0")
AIMessageChunk(
content=" programmation", id="run-9e1517e3-12bf-48f2-bb1b-2e824f7cd7b0"
)
AIMessageChunk(content=".", id="run-9e1517e3-12bf-48f2-bb1b-2e824f7cd7b0")
AIMessageChunk(
content="",
response_metadata={"finish_reason": "stop"},
id="run-9e1517e3-12bf-48f2-bb1b-2e824f7cd7b0",
)
.. code-block:: python
stream = llm.stream(messages)
full = next(stream)
for chunk in stream:
full += chunk
full
.. code-block:: python
AIMessageChunk(
content="J'adore la programmation.",
response_metadata={"finish_reason": "stop"},
id="run-bf917526-7f58-4683-84f7-36a6b671d140",
)
.. dropdown:: Async
.. code-block:: python
await llm.ainvoke(messages)
# stream:
# async for chunk in (await llm.astream(messages))
# batch:
# await llm.abatch([messages])
.. code-block:: python
AIMessage(
content="J'adore la programmation.",
response_metadata={
"token_usage": {
"completion_tokens": 5,
"prompt_tokens": 31,
"total_tokens": 36,
},
"model_name": "gpt-4o",
"system_fingerprint": "fp_43dfabdef1",
"finish_reason": "stop",
"logprobs": None,
},
id="run-012cffe2-5d3d-424d-83b5-51c6d4a593d1-0",
usage_metadata={"input_tokens": 31, "output_tokens": 5, "total_tokens": 36},
)
.. dropdown:: Tool calling
.. code-block:: python
from pydantic import BaseModel, Field
class GetWeather(BaseModel):
'''Get the current weather in a given location'''
location: str = Field(
..., description="The city and state, e.g. San Francisco, CA"
)
class GetPopulation(BaseModel):
'''Get the current population in a given location'''
location: str = Field(
..., description="The city and state, e.g. San Francisco, CA"
)
llm_with_tools = llm.bind_tools(
[GetWeather, GetPopulation]
# strict = True # enforce tool args schema is respected
)
ai_msg = llm_with_tools.invoke(
"Which city is hotter today and which is bigger: LA or NY?"
)
ai_msg.tool_calls
.. code-block:: python
[
{
"name": "GetWeather",
"args": {"location": "Los Angeles, CA"},
"id": "call_6XswGD5Pqk8Tt5atYr7tfenU",
},
{
"name": "GetWeather",
"args": {"location": "New York, NY"},
"id": "call_ZVL15vA8Y7kXqOy3dtmQgeCi",
},
{
"name": "GetPopulation",
"args": {"location": "Los Angeles, CA"},
"id": "call_49CFW8zqC9W7mh7hbMLSIrXw",
},
{
"name": "GetPopulation",
"args": {"location": "New York, NY"},
"id": "call_6ghfKxV264jEfe1mRIkS3PE7",
},
]
Note that ``openai >= 1.32`` supports a ``parallel_tool_calls`` parameter
that defaults to ``True``. This parameter can be set to ``False`` to
disable parallel tool calls:
.. code-block:: python
ai_msg = llm_with_tools.invoke(
"What is the weather in LA and NY?", parallel_tool_calls=False
)
ai_msg.tool_calls
.. code-block:: python
[
{
"name": "GetWeather",
"args": {"location": "Los Angeles, CA"},
"id": "call_4OoY0ZR99iEvC7fevsH8Uhtz",
}
]
Like other runtime parameters, ``parallel_tool_calls`` can be bound to a model
using ``llm.bind(parallel_tool_calls=False)`` or during instantiation by
setting ``model_kwargs``.
See ``ChatOpenAI.bind_tools()`` method for more.
.. dropdown:: Structured output
.. code-block:: python
from typing import Optional
from pydantic import BaseModel, Field
class Joke(BaseModel):
'''Joke to tell user.'''
setup: str = Field(description="The setup of the joke")
punchline: str = Field(description="The punchline to the joke")
rating: Optional[int] = Field(description="How funny the joke is, from 1 to 10")
structured_llm = llm.with_structured_output(Joke)
structured_llm.invoke("Tell me a joke about cats")
.. code-block:: python
Joke(
setup="Why was the cat sitting on the computer?",
punchline="To keep an eye on the mouse!",
rating=None,
)
See ``ChatOpenAI.with_structured_output()`` for more.
.. dropdown:: JSON mode
.. code-block:: python
json_llm = llm.bind(response_format={"type": "json_object"})
ai_msg = json_llm.invoke(
"Return a JSON object with key 'random_ints' and a value of 10 random ints in [0-99]"
)
ai_msg.content
.. code-block:: python
'\\n{\\n "random_ints": [23, 87, 45, 12, 78, 34, 56, 90, 11, 67]\\n}'
.. dropdown:: Image input
.. code-block:: python
import base64
import httpx
from langchain_core.messages import HumanMessage
image_url = "https://upload.wikimedia.org/wikipedia/commons/thumb/d/dd/Gfp-wisconsin-madison-the-nature-boardwalk.jpg/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg"
image_data = base64.b64encode(httpx.get(image_url).content).decode("utf-8")
message = HumanMessage(
content=[
{"type": "text", "text": "describe the weather in this image"},
{
"type": "image_url",
"image_url": {"url": f"data:image/jpeg;base64,{image_data}"},
},
]
)
ai_msg = llm.invoke([message])
ai_msg.content
.. code-block:: python
"The weather in the image appears to be clear and pleasant. The sky is mostly blue with scattered, light clouds, suggesting a sunny day with minimal cloud cover. There is no indication of rain or strong winds, and the overall scene looks bright and calm. The lush green grass and clear visibility further indicate good weather conditions."
.. dropdown:: Token usage
.. code-block:: python
ai_msg = llm.invoke(messages)
ai_msg.usage_metadata
.. code-block:: python
{"input_tokens": 28, "output_tokens": 5, "total_tokens": 33}
When streaming, set the ``stream_usage`` kwarg:
.. code-block:: python
stream = llm.stream(messages, stream_usage=True)
full = next(stream)
for chunk in stream:
full += chunk
full.usage_metadata
.. code-block:: python
{"input_tokens": 28, "output_tokens": 5, "total_tokens": 33}
Alternatively, setting ``stream_usage`` when instantiating the model can be
useful when incorporating ``ChatOpenAI`` into LCEL chains-- or when using
methods like ``.with_structured_output``, which generate chains under the
hood.
.. code-block:: python
llm = ChatOpenAI(model="gpt-4o", stream_usage=True)
structured_llm = llm.with_structured_output(...)
.. dropdown:: Logprobs
.. code-block:: python
logprobs_llm = llm.bind(logprobs=True)
ai_msg = logprobs_llm.invoke(messages)
ai_msg.response_metadata["logprobs"]
.. code-block:: python
{
"content": [
{
"token": "J",
"bytes": [74],
"logprob": -4.9617593e-06,
"top_logprobs": [],
},
{
"token": "'adore",
"bytes": [39, 97, 100, 111, 114, 101],
"logprob": -0.25202933,
"top_logprobs": [],
},
{
"token": " la",
"bytes": [32, 108, 97],
"logprob": -0.20141791,
"top_logprobs": [],
},
{
"token": " programmation",
"bytes": [
32,
112,
114,
111,
103,
114,
97,
109,
109,
97,
116,
105,
111,
110,
],
"logprob": -1.9361265e-07,
"top_logprobs": [],
},
{
"token": ".",
"bytes": [46],
"logprob": -1.2233183e-05,
"top_logprobs": [],
},
]
}
.. dropdown:: Response metadata
.. code-block:: python
ai_msg = llm.invoke(messages)
ai_msg.response_metadata
.. code-block:: python
{
"token_usage": {
"completion_tokens": 5,
"prompt_tokens": 28,
"total_tokens": 33,
},
"model_name": "gpt-4o",
"system_fingerprint": "fp_319be4768e",
"finish_reason": "stop",
"logprobs": None,
}
""" # noqa: E501
stream_usage: bool = False
"""Whether to include usage metadata in streaming output. If True, additional
message chunks will be generated during the stream including usage metadata.
"""
max_tokens: Optional[int] = Field(default=None, alias="max_completion_tokens")
"""Maximum number of tokens to generate."""
@property
def lc_secrets(self) -> Dict[str, str]:
return {"openai_api_key": "OPENAI_API_KEY"}
@classmethod
def get_lc_namespace(cls) -> List[str]:
"""Get the namespace of the langchain object."""
return ["langchain", "chat_models", "openai"]
@property
def lc_attributes(self) -> Dict[str, Any]:
attributes: Dict[str, Any] = {}
if self.openai_organization:
attributes["openai_organization"] = self.openai_organization
if self.openai_api_base:
attributes["openai_api_base"] = self.openai_api_base
if self.openai_proxy:
attributes["openai_proxy"] = self.openai_proxy
return attributes
@classmethod
def is_lc_serializable(cls) -> bool:
"""Return whether this model can be serialized by Langchain."""
return True
@property
def _default_params(self) -> Dict[str, Any]:
"""Get the default parameters for calling OpenAI API."""
params = super()._default_params
if "max_tokens" in params:
params["max_completion_tokens"] = params.pop("max_tokens")
return params
def _get_request_payload(
self,
input_: LanguageModelInput,
*,
stop: Optional[List[str]] = None,
**kwargs: Any,
) -> dict:
payload = super()._get_request_payload(input_, stop=stop, **kwargs)
# max_tokens was deprecated in favor of max_completion_tokens
# in September 2024 release
if "max_tokens" in payload:
payload["max_completion_tokens"] = payload.pop("max_tokens")
# Mutate system message role to "developer" for o-series models
if self.model_name and re.match(r"^o\d", self.model_name):
for message in payload.get("messages", []):
if message["role"] == "system":
message["role"] = "developer"
return payload
def _should_stream_usage(
self, stream_usage: Optional[bool] = None, **kwargs: Any
) -> bool:
"""Determine whether to include usage metadata in streaming output.
For backwards compatibility, we check for `stream_options` passed
explicitly to kwargs or in the model_kwargs and override self.stream_usage.
"""
stream_usage_sources = [ # order of preference
stream_usage,
kwargs.get("stream_options", {}).get("include_usage"),
self.model_kwargs.get("stream_options", {}).get("include_usage"),
self.stream_usage,
]
for source in stream_usage_sources:
if isinstance(source, bool):
return source
return self.stream_usage
def _stream(
self, *args: Any, stream_usage: Optional[bool] = None, **kwargs: Any
) -> Iterator[ChatGenerationChunk]:
"""Set default stream_options."""
stream_usage = self._should_stream_usage(stream_usage, **kwargs)
# Note: stream_options is not a valid parameter for Azure OpenAI.
# To support users proxying Azure through ChatOpenAI, here we only specify
# stream_options if include_usage is set to True.
# See https://learn.microsoft.com/en-us/azure/ai-services/openai/whats-new
# for release notes.
if stream_usage:
kwargs["stream_options"] = {"include_usage": stream_usage}
return super()._stream(*args, **kwargs)
async def _astream(
self, *args: Any, stream_usage: Optional[bool] = None, **kwargs: Any
) -> AsyncIterator[ChatGenerationChunk]:
"""Set default stream_options."""
stream_usage = self._should_stream_usage(stream_usage, **kwargs)
if stream_usage:
kwargs["stream_options"] = {"include_usage": stream_usage}
async for chunk in super()._astream(*args, **kwargs):
yield chunk
def with_structured_output(
self,
schema: Optional[_DictOrPydanticClass] = None,
*,
method: Literal["function_calling", "json_mode", "json_schema"] = "json_schema",
include_raw: bool = False,
strict: Optional[bool] = None,
**kwargs: Any,
) -> Runnable[LanguageModelInput, _DictOrPydantic]:
"""Model wrapper that returns outputs formatted to match the given schema.
Args:
schema:
The output schema. Can be passed in as:
- a JSON Schema,
- a TypedDict class,
- or a Pydantic class,
- an OpenAI function/tool schema.
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, one of:
- "json_schema":
Uses OpenAI's Structured Output API:
https://platform.openai.com/docs/guides/structured-outputs
Supported for "gpt-4o-mini", "gpt-4o-2024-08-06", "o1", and later
models.
- "function_calling":
Uses OpenAI's tool-calling (formerly called function calling)
API: https://platform.openai.com/docs/guides/function-calling
- "json_mode":
Uses OpenAI's JSON mode. Note that if using JSON mode then you
must include instructions for formatting the output into the
desired schema into the model call:
https://platform.openai.com/docs/guides/structured-outputs/json-mode
Learn more about the differences between the methods and which models
support which methods here:
- https://platform.openai.com/docs/guides/structured-outputs/structured-outputs-vs-json-mode
- https://platform.openai.com/docs/guides/structured-outputs/function-calling-vs-response-format
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".
strict:
- True:
Model output is guaranteed to exactly match the schema.
The input schema will also be validated according to
https://platform.openai.com/docs/guides/structured-outputs/supported-schemas
- False:
Input schema will not be validated and model output will not be
validated.
- None:
``strict`` argument will not be passed to the model.
If schema is specified via TypedDict or JSON schema, ``strict`` is not
enabled by default. Pass ``strict=True`` to enable it.
Note: ``strict`` can only be non-null if ``method`` is
``"json_schema"`` or ``"function_calling"``.
kwargs: Additional keyword args aren't supported.
Returns:
A Runnable that takes same inputs as a :class:`langchain_core.language_models.chat.BaseChatModel`.
| If ``include_raw`` is False and ``schema`` is a Pydantic class, Runnable outputs an instance of ``schema`` (i.e., a Pydantic object). Otherwise, if ``include_raw`` is False then Runnable outputs a dict.
| 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]
.. versionchanged:: 0.1.20
Added support for TypedDict class ``schema``.
.. versionchanged:: 0.1.21
Support for ``strict`` argument added.
Support for ``method="json_schema"`` added.
.. versionchanged:: 0.3.0
``method`` default changed from "function_calling" to "json_schema".
.. dropdown:: Example: schema=Pydantic class, method="json_schema", include_raw=False, strict=True
Note, OpenAI has a number of restrictions on what types of schemas can be
provided if ``strict`` = True. When using Pydantic, our model cannot
specify any Field metadata (like min/max constraints) and fields cannot
have default values.
See all constraints here: https://platform.openai.com/docs/guides/structured-outputs/supported-schemas
.. code-block:: python
from typing import Optional
from langchain_openai import ChatOpenAI
from pydantic import BaseModel, Field
class AnswerWithJustification(BaseModel):
'''An answer to the user question along with justification for the answer.'''
answer: str
justification: Optional[str] = Field(
default=..., description="A justification for the answer."
)
llm = ChatOpenAI(model="gpt-4o", temperature=0)
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.'
# )
.. dropdown:: Example: schema=Pydantic class, method="function_calling", include_raw=False, strict=False
.. code-block:: python
from typing import Optional
from langchain_openai import ChatOpenAI
from pydantic import BaseModel, Field
class AnswerWithJustification(BaseModel):
'''An answer to the user question along with justification for the answer.'''
answer: str
justification: Optional[str] = Field(
default=..., description="A justification for the answer."
)
llm = ChatOpenAI(model="gpt-4o", temperature=0)
structured_llm = llm.with_structured_output(
AnswerWithJustification, method="function_calling"
)
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.'
# )
.. dropdown:: Example: schema=Pydantic class, method="json_schema", include_raw=True
.. code-block:: python
from langchain_openai import ChatOpenAI
from pydantic import BaseModel
class AnswerWithJustification(BaseModel):
'''An answer to the user question along with justification for the answer.'''
answer: str
justification: str
llm = ChatOpenAI(model="gpt-4o", temperature=0)
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
# }
.. dropdown:: Example: schema=TypedDict class, method="json_schema", include_raw=False, strict=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
class AnswerWithJustification(TypedDict):
'''An answer to the user question along with justification for the answer.'''
answer: str
justification: Annotated[
Optional[str], None, "A justification for the answer."
]
llm = ChatOpenAI(model="gpt-4o", temperature=0)
structured_llm = llm.with_structured_output(AnswerWithJustification)
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.'
# }
.. dropdown:: Example: schema=OpenAI function schema, method="json_schema", 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-4o", 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.'
# }
.. dropdown:: Example: schema=Pydantic class, method="json_mode", include_raw=True
.. code-block::
from langchain_openai import ChatOpenAI
from pydantic import BaseModel
class AnswerWithJustification(BaseModel):
answer: str
justification: str
llm = ChatOpenAI(model="gpt-4o", temperature=0)
structured_llm = llm.with_structured_output(
AnswerWithJustification,
method="json_mode",
include_raw=True
)
structured_llm.invoke(
"Answer the following question. "
"Make sure to return a JSON blob with keys 'answer' and 'justification'.\\n\\n"
"What's heavier a pound of bricks or a pound of feathers?"
)
# -> {
# 'raw': AIMessage(content='{\\n "answer": "They are both the same weight.",\\n "justification": "Both a pound of bricks and a pound of feathers weigh one pound. The difference lies in the volume and density of the materials, not the weight." \\n}'),
# 'parsed': AnswerWithJustification(answer='They are both the same weight.', justification='Both a pound of bricks and a pound of feathers weigh one pound. The difference lies in the volume and density of the materials, not the weight.'),
# 'parsing_error': None
# }
.. dropdown:: Example: schema=None, method="json_mode", include_raw=True
.. code-block::
structured_llm = llm.with_structured_output(method="json_mode", include_raw=True)
structured_llm.invoke(
"Answer the following question. "
"Make sure to return a JSON blob with keys 'answer' and 'justification'.\\n\\n"
"What's heavier a pound of bricks or a pound of feathers?"
)
# -> {
# 'raw': AIMessage(content='{\\n "answer": "They are both the same weight.",\\n "justification": "Both a pound of bricks and a pound of feathers weigh one pound. The difference lies in the volume and density of the materials, not the weight." \\n}'),
# 'parsed': {
# 'answer': 'They are both the same weight.',
# 'justification': 'Both a pound of bricks and a pound of feathers weigh one pound. The difference lies in the volume and density of the materials, not the weight.'
# },
# 'parsing_error': None
# }
""" # noqa: E501
return super().with_structured_output(
schema, method=method, include_raw=include_raw, strict=strict, **kwargs
)
def _is_pydantic_class(obj: Any) -> bool:
return isinstance(obj, type) and is_basemodel_subclass(obj)
def _lc_tool_call_to_openai_tool_call(tool_call: ToolCall) -> dict:
return {
"type": "function",
"id": tool_call["id"],
"function": {
"name": tool_call["name"],
"arguments": json.dumps(tool_call["args"]),
},
}
def _lc_invalid_tool_call_to_openai_tool_call(
invalid_tool_call: InvalidToolCall,
) -> dict:
return {
"type": "function",
"id": invalid_tool_call["id"],
"function": {
"name": invalid_tool_call["name"],
"arguments": invalid_tool_call["args"],
},
}
def _url_to_size(image_source: str) -> Optional[Tuple[int, int]]:
try:
from PIL import Image # type: ignore[import]
except ImportError:
logger.info(
"Unable to count image tokens. To count image tokens please install "
"`pip install -U pillow httpx`."
)
return None
if _is_url(image_source):
try:
import httpx
except ImportError:
logger.info(
"Unable to count image tokens. To count image tokens please install "
"`pip install -U httpx`."
)
return None
response = httpx.get(image_source)
response.raise_for_status()
width, height = Image.open(BytesIO(response.content)).size
return width, height
elif _is_b64(image_source):
_, encoded = image_source.split(",", 1)
data = base64.b64decode(encoded)
width, height = Image.open(BytesIO(data)).size
return width, height
else:
return None
def _count_image_tokens(width: int, height: int) -> int:
# Reference: https://platform.openai.com/docs/guides/vision/calculating-costs
width, height = _resize(width, height)
h = ceil(height / 512)
w = ceil(width / 512)
return (170 * h * w) + 85
def _is_url(s: str) -> bool:
try:
result = urlparse(s)
return all([result.scheme, result.netloc])
except Exception as e:
logger.debug(f"Unable to parse URL: {e}")
return False
def _is_b64(s: str) -> bool:
return s.startswith("data:image")
def _resize(width: int, height: int) -> Tuple[int, int]:
# larger side must be <= 2048
if width > 2048 or height > 2048:
if width > height:
height = (height * 2048) // width
width = 2048
else:
width = (width * 2048) // height
height = 2048
# smaller side must be <= 768
if width > 768 and height > 768:
if width > height:
width = (width * 768) // height
height = 768
else:
height = (width * 768) // height
width = 768
return width, height
def _convert_to_openai_response_format(
schema: Union[Dict[str, Any], Type], *, strict: Optional[bool] = None
) -> Union[Dict, TypeBaseModel]:
if isinstance(schema, type) and is_basemodel_subclass(schema):
return schema
if (
isinstance(schema, dict)
and "json_schema" in schema
and schema.get("type") == "json_schema"
):
response_format = schema
elif isinstance(schema, dict) and "name" in schema and "schema" in schema:
response_format = {"type": "json_schema", "json_schema": schema}
else:
if strict is None:
if isinstance(schema, dict) and isinstance(schema.get("strict"), bool):
strict = schema["strict"]
else:
strict = False
function = convert_to_openai_tool(schema, strict=strict)["function"]
function["schema"] = function.pop("parameters")
response_format = {"type": "json_schema", "json_schema": function}
if strict is not None and strict is not response_format["json_schema"].get(
"strict"
):
msg = (
f"Output schema already has 'strict' value set to "
f"{schema['json_schema']['strict']} but 'strict' also passed in to "
f"with_structured_output as {strict}. Please make sure that "
f"'strict' is only specified in one place."
)
raise ValueError(msg)
return response_format
def _oai_structured_outputs_parser(
ai_msg: AIMessage, schema: Type[_BM]
) -> PydanticBaseModel:
if parsed := ai_msg.additional_kwargs.get("parsed"):
if isinstance(parsed, dict):
return schema(**parsed)
else:
return parsed
elif ai_msg.additional_kwargs.get("refusal"):
raise OpenAIRefusalError(ai_msg.additional_kwargs["refusal"])
else:
raise ValueError(
"Structured Output response does not have a 'parsed' field nor a 'refusal' "
f"field. Received message:\n\n{ai_msg}"
)
class OpenAIRefusalError(Exception):
"""Error raised when OpenAI Structured Outputs API returns a refusal.
When using OpenAI's Structured Outputs API with user-generated input, the model
may occasionally refuse to fulfill the request for safety reasons.
See here for more on refusals:
https://platform.openai.com/docs/guides/structured-outputs/refusals
.. versionadded:: 0.1.21
"""
def _create_usage_metadata(oai_token_usage: dict) -> UsageMetadata:
input_tokens = oai_token_usage.get("prompt_tokens", 0)
output_tokens = oai_token_usage.get("completion_tokens", 0)
total_tokens = oai_token_usage.get("total_tokens", input_tokens + output_tokens)
input_token_details: dict = {
"audio": (oai_token_usage.get("prompt_tokens_details") or {}).get(
"audio_tokens"
),
"cache_read": (oai_token_usage.get("prompt_tokens_details") or {}).get(
"cached_tokens"
),
}
output_token_details: dict = {
"audio": (oai_token_usage.get("completion_tokens_details") or {}).get(
"audio_tokens"
),
"reasoning": (oai_token_usage.get("completion_tokens_details") or {}).get(
"reasoning_tokens"
),
}
return UsageMetadata(
input_tokens=input_tokens,
output_tokens=output_tokens,
total_tokens=total_tokens,
input_token_details=InputTokenDetails(
**{k: v for k, v in input_token_details.items() if v is not None}
),
output_token_details=OutputTokenDetails(
**{k: v for k, v in output_token_details.items() if v is not None}
),
)
def _convert_to_json_schema(
schema: Union[Dict[str, Any], Type], *, strict: Optional[bool] = None
) -> Union[Dict, TypeBaseModel]:
if _is_pydantic_class(schema):
schema = convert_to_openai_function(schema)
response_format = cast(
dict, _convert_to_openai_response_format(schema, strict=strict)
)
json_schema = response_format["json_schema"]["schema"]
if "name" in response_format["json_schema"]:
json_schema["title"] = response_format["json_schema"]["name"]
if "description" in response_format["json_schema"]:
json_schema["description"] = response_format["json_schema"]["description"]
return json_schema
Reference in New Issue View Git Blame Copy Permalink