mirror of
https://github.com/hwchase17/langchain
synced 2024-11-11 19:11:02 +00:00
8461934c2b
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
1032 lines
42 KiB
Python
1032 lines
42 KiB
Python
"""Fireworks chat wrapper."""
|
|
|
|
from __future__ import annotations
|
|
|
|
import json
|
|
import logging
|
|
import os
|
|
from operator import itemgetter
|
|
from typing import (
|
|
Any,
|
|
AsyncIterator,
|
|
Callable,
|
|
Dict,
|
|
Iterator,
|
|
List,
|
|
Literal,
|
|
Mapping,
|
|
Optional,
|
|
Sequence,
|
|
Tuple,
|
|
Type,
|
|
TypedDict,
|
|
Union,
|
|
cast,
|
|
)
|
|
|
|
from fireworks.client import AsyncFireworks, Fireworks # type: ignore
|
|
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.tool import (
|
|
ToolCallChunk,
|
|
)
|
|
from langchain_core.messages.tool import (
|
|
tool_call_chunk as create_tool_call_chunk,
|
|
)
|
|
from langchain_core.output_parsers import JsonOutputParser, PydanticOutputParser
|
|
from langchain_core.output_parsers.base import OutputParserLike
|
|
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.pydantic_v1 import (
|
|
BaseModel,
|
|
Field,
|
|
SecretStr,
|
|
root_validator,
|
|
)
|
|
from langchain_core.runnables import Runnable, RunnableMap, RunnablePassthrough
|
|
from langchain_core.tools import BaseTool
|
|
from langchain_core.utils import (
|
|
convert_to_secret_str,
|
|
get_from_dict_or_env,
|
|
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 is_basemodel_subclass
|
|
from langchain_core.utils.utils import build_extra_kwargs
|
|
|
|
logger = logging.getLogger(__name__)
|
|
|
|
|
|
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")
|
|
if role == "user":
|
|
return HumanMessage(content=_dict.get("content", ""))
|
|
elif role == "assistant":
|
|
# Fix for azure
|
|
# Also Fireworks 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(
|
|
dict(make_invalid_tool_call(raw_tool_call, str(e)))
|
|
)
|
|
return AIMessage(
|
|
content=content,
|
|
additional_kwargs=additional_kwargs,
|
|
tool_calls=tool_calls,
|
|
invalid_tool_calls=invalid_tool_calls,
|
|
)
|
|
elif role == "system":
|
|
return SystemMessage(content=_dict.get("content", ""))
|
|
elif role == "function":
|
|
return FunctionMessage(
|
|
content=_dict.get("content", ""), name=_dict.get("name", "")
|
|
)
|
|
elif role == "tool":
|
|
additional_kwargs = {}
|
|
if "name" in _dict:
|
|
additional_kwargs["name"] = _dict["name"]
|
|
return ToolMessage(
|
|
content=_dict.get("content", ""),
|
|
tool_call_id=_dict.get("tool_call_id", ""),
|
|
additional_kwargs=additional_kwargs,
|
|
)
|
|
else:
|
|
return ChatMessage(content=_dict.get("content", ""), role=role or "")
|
|
|
|
|
|
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]
|
|
if isinstance(message, ChatMessage):
|
|
message_dict = {"role": message.role, "content": message.content}
|
|
elif isinstance(message, HumanMessage):
|
|
message_dict = {"role": "user", "content": message.content}
|
|
elif isinstance(message, AIMessage):
|
|
message_dict = {"role": "assistant", "content": message.content}
|
|
if "function_call" in message.additional_kwargs:
|
|
message_dict["function_call"] = message.additional_kwargs["function_call"]
|
|
# If function call only, content is None not empty string
|
|
if message_dict["content"] == "":
|
|
message_dict["content"] = None
|
|
if message.tool_calls or message.invalid_tool_calls:
|
|
message_dict["tool_calls"] = [
|
|
_lc_tool_call_to_fireworks_tool_call(tc) for tc in message.tool_calls
|
|
] + [
|
|
_lc_invalid_tool_call_to_fireworks_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"]
|
|
# If tool calls only, content is None not empty string
|
|
if "tool_calls" in message_dict and message_dict["content"] == "":
|
|
message_dict["content"] = None
|
|
else:
|
|
pass
|
|
elif isinstance(message, SystemMessage):
|
|
message_dict = {"role": "system", "content": message.content}
|
|
elif isinstance(message, FunctionMessage):
|
|
message_dict = {
|
|
"role": "function",
|
|
"content": message.content,
|
|
"name": message.name,
|
|
}
|
|
elif isinstance(message, ToolMessage):
|
|
message_dict = {
|
|
"role": "tool",
|
|
"content": message.content,
|
|
"tool_call_id": message.tool_call_id,
|
|
}
|
|
else:
|
|
raise TypeError(f"Got unknown type {message}")
|
|
if "name" in message.additional_kwargs:
|
|
message_dict["name"] = message.additional_kwargs["name"]
|
|
return message_dict
|
|
|
|
|
|
def _convert_chunk_to_message_chunk(
|
|
chunk: Mapping[str, Any], default_class: Type[BaseMessageChunk]
|
|
) -> BaseMessageChunk:
|
|
choice = chunk["choices"][0]
|
|
_dict = choice["delta"]
|
|
role = cast(str, _dict.get("role"))
|
|
content = cast(str, _dict.get("content") or "")
|
|
additional_kwargs: Dict = {}
|
|
tool_call_chunks: List[ToolCallChunk] = []
|
|
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
|
|
if raw_tool_calls := _dict.get("tool_calls"):
|
|
additional_kwargs["tool_calls"] = raw_tool_calls
|
|
for rtc in raw_tool_calls:
|
|
try:
|
|
tool_call_chunks.append(
|
|
create_tool_call_chunk(
|
|
name=rtc["function"].get("name"),
|
|
args=rtc["function"].get("arguments"),
|
|
id=rtc.get("id"),
|
|
index=rtc.get("index"),
|
|
)
|
|
)
|
|
except KeyError:
|
|
pass
|
|
if role == "user" or default_class == HumanMessageChunk:
|
|
return HumanMessageChunk(content=content)
|
|
elif role == "assistant" or default_class == AIMessageChunk:
|
|
if usage := chunk.get("usage"):
|
|
input_tokens = usage.get("prompt_tokens", 0)
|
|
output_tokens = usage.get("completion_tokens", 0)
|
|
usage_metadata = {
|
|
"input_tokens": input_tokens,
|
|
"output_tokens": output_tokens,
|
|
"total_tokens": usage.get("total_tokens", input_tokens + output_tokens),
|
|
}
|
|
else:
|
|
usage_metadata = None
|
|
return AIMessageChunk(
|
|
content=content,
|
|
additional_kwargs=additional_kwargs,
|
|
tool_call_chunks=tool_call_chunks,
|
|
usage_metadata=usage_metadata, # type: ignore[arg-type]
|
|
)
|
|
elif role == "system" or default_class == SystemMessageChunk:
|
|
return SystemMessageChunk(content=content)
|
|
elif role == "function" or default_class == FunctionMessageChunk:
|
|
return FunctionMessageChunk(content=content, name=_dict["name"])
|
|
elif role == "tool" or default_class == ToolMessageChunk:
|
|
return ToolMessageChunk(content=content, tool_call_id=_dict["tool_call_id"])
|
|
elif role or default_class == ChatMessageChunk:
|
|
return ChatMessageChunk(content=content, role=role)
|
|
else:
|
|
return default_class(content=content) # type: ignore
|
|
|
|
|
|
class _FunctionCall(TypedDict):
|
|
name: str
|
|
|
|
|
|
# This is basically a copy and replace for ChatFireworks, except
|
|
# - I needed to gut out tiktoken and some of the token estimation logic
|
|
# (not sure how important it is)
|
|
# - Environment variable is different
|
|
# we should refactor into some OpenAI-like class in the future
|
|
class ChatFireworks(BaseChatModel):
|
|
"""`Fireworks` Chat large language models API.
|
|
|
|
To use, you should have the
|
|
environment variable ``FIREWORKS_API_KEY`` set with your API key.
|
|
|
|
Any parameters that are valid to be passed to the fireworks.create call
|
|
can be passed in, even if not explicitly saved on this class.
|
|
|
|
Example:
|
|
.. code-block:: python
|
|
|
|
from langchain_fireworks.chat_models import ChatFireworks
|
|
fireworks = ChatFireworks(
|
|
model_name="accounts/fireworks/models/mixtral-8x7b-instruct")
|
|
"""
|
|
|
|
@property
|
|
def lc_secrets(self) -> Dict[str, str]:
|
|
return {"fireworks_api_key": "FIREWORKS_API_KEY"}
|
|
|
|
@classmethod
|
|
def get_lc_namespace(cls) -> List[str]:
|
|
"""Get the namespace of the langchain object."""
|
|
return ["langchain", "chat_models", "fireworks"]
|
|
|
|
@property
|
|
def lc_attributes(self) -> Dict[str, Any]:
|
|
attributes: Dict[str, Any] = {}
|
|
if self.fireworks_api_base:
|
|
attributes["fireworks_api_base"] = self.fireworks_api_base
|
|
|
|
return attributes
|
|
|
|
@classmethod
|
|
def is_lc_serializable(cls) -> bool:
|
|
"""Return whether this model can be serialized by Langchain."""
|
|
return True
|
|
|
|
client: Any = Field(default=None, exclude=True) #: :meta private:
|
|
async_client: Any = Field(default=None, exclude=True) #: :meta private:
|
|
model_name: str = Field(
|
|
default="accounts/fireworks/models/mixtral-8x7b-instruct", alias="model"
|
|
)
|
|
"""Model name to use."""
|
|
temperature: float = 0.0
|
|
"""What sampling temperature to use."""
|
|
stop: Optional[Union[str, List[str]]] = Field(default=None, alias="stop_sequences")
|
|
"""Default stop sequences."""
|
|
model_kwargs: Dict[str, Any] = Field(default_factory=dict)
|
|
"""Holds any model parameters valid for `create` call not explicitly specified."""
|
|
fireworks_api_key: SecretStr = Field(default=None, alias="api_key")
|
|
"""Automatically inferred from env var `FIREWORKS_API_KEY` if not provided."""
|
|
fireworks_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."""
|
|
request_timeout: Union[float, Tuple[float, float], Any, None] = Field(
|
|
default=None, alias="timeout"
|
|
)
|
|
"""Timeout for requests to Fireworks completion API. Can be float, httpx.Timeout or
|
|
None."""
|
|
streaming: bool = False
|
|
"""Whether to stream the results or not."""
|
|
n: int = 1
|
|
"""Number of chat completions to generate for each prompt."""
|
|
max_tokens: Optional[int] = None
|
|
"""Maximum number of tokens to generate."""
|
|
max_retries: Optional[int] = None
|
|
"""Maximum number of retries to make when generating."""
|
|
|
|
class Config:
|
|
"""Configuration for this pydantic object."""
|
|
|
|
allow_population_by_field_name = True
|
|
|
|
@root_validator(pre=True)
|
|
def build_extra(cls, values: Dict[str, Any]) -> Dict[str, Any]:
|
|
"""Build extra kwargs from additional params that were passed in."""
|
|
all_required_field_names = get_pydantic_field_names(cls)
|
|
extra = values.get("model_kwargs", {})
|
|
values["model_kwargs"] = build_extra_kwargs(
|
|
extra, values, all_required_field_names
|
|
)
|
|
return values
|
|
|
|
@root_validator()
|
|
def validate_environment(cls, values: Dict) -> Dict:
|
|
"""Validate that api key and python package exists in environment."""
|
|
if values["n"] < 1:
|
|
raise ValueError("n must be at least 1.")
|
|
if values["n"] > 1 and values["streaming"]:
|
|
raise ValueError("n must be 1 when streaming.")
|
|
|
|
values["fireworks_api_key"] = convert_to_secret_str(
|
|
get_from_dict_or_env(values, "fireworks_api_key", "FIREWORKS_API_KEY")
|
|
)
|
|
values["fireworks_api_base"] = values["fireworks_api_base"] or os.getenv(
|
|
"FIREWORKS_API_BASE"
|
|
)
|
|
client_params = {
|
|
"api_key": (
|
|
values["fireworks_api_key"].get_secret_value()
|
|
if values["fireworks_api_key"]
|
|
else None
|
|
),
|
|
"base_url": values["fireworks_api_base"],
|
|
"timeout": values["request_timeout"],
|
|
}
|
|
|
|
if not values.get("client"):
|
|
values["client"] = Fireworks(**client_params).chat.completions
|
|
if not values.get("async_client"):
|
|
values["async_client"] = AsyncFireworks(**client_params).chat.completions
|
|
if values["max_retries"]:
|
|
values["client"]._max_retries = values["max_retries"]
|
|
values["async_client"]._max_retries = values["max_retries"]
|
|
return values
|
|
|
|
@property
|
|
def _default_params(self) -> Dict[str, Any]:
|
|
"""Get the default parameters for calling Fireworks API."""
|
|
params = {
|
|
"model": self.model_name,
|
|
"stream": self.streaming,
|
|
"n": self.n,
|
|
"temperature": self.temperature,
|
|
"stop": self.stop,
|
|
**self.model_kwargs,
|
|
}
|
|
if self.max_tokens is not None:
|
|
params["max_tokens"] = self.max_tokens
|
|
return params
|
|
|
|
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="fireworks",
|
|
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):
|
|
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
|
|
|
|
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 k in overall_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 _stream(
|
|
self,
|
|
messages: List[BaseMessage],
|
|
stop: Optional[List[str]] = None,
|
|
run_manager: Optional[CallbackManagerForLLMRun] = None,
|
|
**kwargs: Any,
|
|
) -> Iterator[ChatGenerationChunk]:
|
|
message_dicts, params = self._create_message_dicts(messages, stop)
|
|
params = {**params, **kwargs, "stream": True}
|
|
|
|
default_chunk_class: Type[BaseMessageChunk] = AIMessageChunk
|
|
for chunk in self.client.create(messages=message_dicts, **params):
|
|
if not isinstance(chunk, dict):
|
|
chunk = chunk.dict()
|
|
if len(chunk["choices"]) == 0:
|
|
continue
|
|
choice = chunk["choices"][0]
|
|
message_chunk = _convert_chunk_to_message_chunk(chunk, default_chunk_class)
|
|
generation_info = {}
|
|
if finish_reason := choice.get("finish_reason"):
|
|
generation_info["finish_reason"] = finish_reason
|
|
logprobs = choice.get("logprobs")
|
|
if logprobs:
|
|
generation_info["logprobs"] = logprobs
|
|
default_chunk_class = message_chunk.__class__
|
|
generation_chunk = ChatGenerationChunk(
|
|
message=message_chunk, generation_info=generation_info or None
|
|
)
|
|
if run_manager:
|
|
run_manager.on_llm_new_token(
|
|
generation_chunk.text, chunk=generation_chunk, logprobs=logprobs
|
|
)
|
|
yield generation_chunk
|
|
|
|
def _generate(
|
|
self,
|
|
messages: List[BaseMessage],
|
|
stop: Optional[List[str]] = None,
|
|
run_manager: Optional[CallbackManagerForLLMRun] = None,
|
|
stream: Optional[bool] = None,
|
|
**kwargs: Any,
|
|
) -> ChatResult:
|
|
should_stream = stream if stream is not None else self.streaming
|
|
if should_stream:
|
|
stream_iter = self._stream(
|
|
messages, stop=stop, run_manager=run_manager, **kwargs
|
|
)
|
|
return generate_from_stream(stream_iter)
|
|
message_dicts, params = self._create_message_dicts(messages, stop)
|
|
params = {
|
|
**params,
|
|
**({"stream": stream} if stream is not None else {}),
|
|
**kwargs,
|
|
}
|
|
response = self.client.create(messages=message_dicts, **params)
|
|
return self._create_chat_result(response)
|
|
|
|
def _create_message_dicts(
|
|
self, messages: List[BaseMessage], stop: Optional[List[str]]
|
|
) -> Tuple[List[Dict[str, Any]], Dict[str, Any]]:
|
|
params = self._default_params
|
|
if stop is not None:
|
|
params["stop"] = stop
|
|
message_dicts = [_convert_message_to_dict(m) for m in messages]
|
|
return message_dicts, params
|
|
|
|
def _create_chat_result(self, response: Union[dict, BaseModel]) -> ChatResult:
|
|
generations = []
|
|
if not isinstance(response, dict):
|
|
response = response.dict()
|
|
token_usage = response.get("usage", {})
|
|
for res in response["choices"]:
|
|
message = _convert_dict_to_message(res["message"])
|
|
if token_usage and isinstance(message, AIMessage):
|
|
message.usage_metadata = {
|
|
"input_tokens": token_usage.get("prompt_tokens", 0),
|
|
"output_tokens": token_usage.get("completion_tokens", 0),
|
|
"total_tokens": token_usage.get("total_tokens", 0),
|
|
}
|
|
generation_info = dict(finish_reason=res.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": self.model_name,
|
|
"system_fingerprint": response.get("system_fingerprint", ""),
|
|
}
|
|
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]:
|
|
message_dicts, params = self._create_message_dicts(messages, stop)
|
|
params = {**params, **kwargs, "stream": True}
|
|
|
|
default_chunk_class: Type[BaseMessageChunk] = AIMessageChunk
|
|
async for chunk in self.async_client.acreate(messages=message_dicts, **params):
|
|
if not isinstance(chunk, dict):
|
|
chunk = chunk.dict()
|
|
if len(chunk["choices"]) == 0:
|
|
continue
|
|
choice = chunk["choices"][0]
|
|
message_chunk = _convert_chunk_to_message_chunk(chunk, default_chunk_class)
|
|
generation_info = {}
|
|
if finish_reason := choice.get("finish_reason"):
|
|
generation_info["finish_reason"] = finish_reason
|
|
logprobs = choice.get("logprobs")
|
|
if logprobs:
|
|
generation_info["logprobs"] = logprobs
|
|
default_chunk_class = message_chunk.__class__
|
|
generation_chunk = ChatGenerationChunk(
|
|
message=message_chunk, generation_info=generation_info or None
|
|
)
|
|
if run_manager:
|
|
await run_manager.on_llm_new_token(
|
|
token=generation_chunk.text,
|
|
chunk=generation_chunk,
|
|
logprobs=logprobs,
|
|
)
|
|
yield generation_chunk
|
|
|
|
async def _agenerate(
|
|
self,
|
|
messages: List[BaseMessage],
|
|
stop: Optional[List[str]] = None,
|
|
run_manager: Optional[AsyncCallbackManagerForLLMRun] = None,
|
|
stream: Optional[bool] = None,
|
|
**kwargs: Any,
|
|
) -> ChatResult:
|
|
should_stream = stream if stream is not None else self.streaming
|
|
if should_stream:
|
|
stream_iter = self._astream(
|
|
messages, stop=stop, run_manager=run_manager, **kwargs
|
|
)
|
|
return await agenerate_from_stream(stream_iter)
|
|
|
|
message_dicts, params = self._create_message_dicts(messages, stop)
|
|
params = {
|
|
**params,
|
|
**({"stream": stream} if stream is not None else {}),
|
|
**kwargs,
|
|
}
|
|
response = await self.async_client.acreate(messages=message_dicts, **params)
|
|
return self._create_chat_result(response)
|
|
|
|
@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,
|
|
}
|
|
|
|
@property
|
|
def _llm_type(self) -> str:
|
|
"""Return type of chat model."""
|
|
return "fireworks-chat"
|
|
|
|
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 Fireworks function-calling API.
|
|
|
|
NOTE: Using bind_tools is recommended instead, as the `functions` and
|
|
`function_call` request parameters are officially marked as deprecated by
|
|
Fireworks.
|
|
|
|
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[BaseModel], Callable, BaseTool]],
|
|
*,
|
|
tool_choice: Optional[
|
|
Union[dict, str, Literal["auto", "any", "none"], bool]
|
|
] = None,
|
|
**kwargs: Any,
|
|
) -> Runnable[LanguageModelInput, BaseMessage]:
|
|
"""Bind tool-like objects to this chat model.
|
|
|
|
Assumes model is compatible with Fireworks 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.
|
|
Must be the name of the single provided function,
|
|
"auto" to automatically determine which function to call
|
|
with the option to not call any function, "any" to enforce that some
|
|
function is called, 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.
|
|
"""
|
|
|
|
formatted_tools = [convert_to_openai_tool(tool) for tool in tools]
|
|
if tool_choice is not None and tool_choice:
|
|
if isinstance(tool_choice, str) and (
|
|
tool_choice not in ("auto", "any", "none")
|
|
):
|
|
tool_choice = {"type": "function", "function": {"name": tool_choice}}
|
|
if isinstance(tool_choice, dict) and (len(formatted_tools) != 1):
|
|
raise ValueError(
|
|
"When specifying `tool_choice`, you must provide exactly one "
|
|
f"tool. Received {len(formatted_tools)} tools."
|
|
)
|
|
if isinstance(tool_choice, dict) and (
|
|
formatted_tools[0]["function"]["name"]
|
|
!= tool_choice["function"]["name"]
|
|
):
|
|
raise ValueError(
|
|
f"Tool choice {tool_choice} was specified, but the only "
|
|
f"provided tool was {formatted_tools[0]['function']['name']}."
|
|
)
|
|
if isinstance(tool_choice, bool):
|
|
if len(tools) > 1:
|
|
raise ValueError(
|
|
"tool_choice can only be True when there is one tool. Received "
|
|
f"{len(tools)} tools."
|
|
)
|
|
tool_name = formatted_tools[0]["function"]["name"]
|
|
tool_choice = {
|
|
"type": "function",
|
|
"function": {"name": tool_name},
|
|
}
|
|
|
|
kwargs["tool_choice"] = tool_choice
|
|
return super().bind(tools=formatted_tools, **kwargs)
|
|
|
|
def with_structured_output(
|
|
self,
|
|
schema: Optional[Union[Dict, Type[BaseModel]]] = None,
|
|
*,
|
|
method: Literal["function_calling", "json_mode"] = "function_calling",
|
|
include_raw: bool = False,
|
|
**kwargs: Any,
|
|
) -> Runnable[LanguageModelInput, Union[Dict, BaseModel]]:
|
|
"""Model wrapper that returns outputs formatted to match the given schema.
|
|
|
|
Args:
|
|
schema:
|
|
The output schema. Can be passed in as:
|
|
- an OpenAI function/tool schema,
|
|
- a JSON Schema,
|
|
- a TypedDict class (support added in 0.1.7),
|
|
- 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:: 0.1.7
|
|
|
|
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
|
|
an error occurs during model output parsing it will be raised. If True
|
|
then both the raw model response (a BaseMessage) and the parsed model
|
|
response will be returned. If an error occurs during output parsing it
|
|
will be caught and returned as well. The final output is always a dict
|
|
with keys "raw", "parsed", and "parsing_error".
|
|
|
|
Returns:
|
|
A Runnable that takes 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]
|
|
|
|
Example: schema=Pydantic class, method="function_calling", include_raw=False:
|
|
.. code-block:: python
|
|
|
|
from typing import Optional
|
|
|
|
from langchain_fireworks import ChatFireworks
|
|
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
|
|
# 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 = ChatFireworks(model="accounts/fireworks/models/firefunction-v1", 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.'
|
|
# )
|
|
|
|
Example: schema=Pydantic class, method="function_calling", include_raw=True:
|
|
.. code-block:: python
|
|
|
|
from langchain_fireworks import ChatFireworks
|
|
from langchain_core.pydantic_v1 import BaseModel
|
|
|
|
|
|
class AnswerWithJustification(BaseModel):
|
|
'''An answer to the user question along with justification for the answer.'''
|
|
|
|
answer: str
|
|
justification: str
|
|
|
|
|
|
llm = ChatFireworks(model="accounts/fireworks/models/firefunction-v1", 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
|
|
# }
|
|
|
|
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_fireworks import ChatFireworks
|
|
|
|
|
|
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 = ChatFireworks(model="accounts/fireworks/models/firefunction-v1", 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.'
|
|
# }
|
|
|
|
Example: schema=OpenAI function schema, method="function_calling", include_raw=False:
|
|
.. code-block:: python
|
|
|
|
from langchain_fireworks import ChatFireworks
|
|
|
|
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 = ChatFireworks(model="accounts/fireworks/models/firefunction-v1", 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_fireworks import ChatFireworks
|
|
from langchain_core.pydantic_v1 import BaseModel
|
|
|
|
class AnswerWithJustification(BaseModel):
|
|
answer: str
|
|
justification: str
|
|
|
|
llm = ChatFireworks(model="accounts/fireworks/models/firefunction-v1", 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
|
|
# }
|
|
|
|
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
|
|
if kwargs:
|
|
raise ValueError(f"Received unsupported arguments {kwargs}")
|
|
is_pydantic_schema = _is_pydantic_class(schema)
|
|
if method == "function_calling":
|
|
if schema is None:
|
|
raise ValueError(
|
|
"schema must be specified when method is 'function_calling'. "
|
|
"Received None."
|
|
)
|
|
tool_name = convert_to_openai_tool(schema)["function"]["name"]
|
|
llm = self.bind_tools([schema], tool_choice=tool_name)
|
|
if is_pydantic_schema:
|
|
output_parser: OutputParserLike = 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"})
|
|
output_parser = (
|
|
PydanticOutputParser(pydantic_object=schema) # type: ignore[type-var, arg-type]
|
|
if is_pydantic_schema
|
|
else 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 _is_pydantic_class(obj: Any) -> bool:
|
|
return isinstance(obj, type) and is_basemodel_subclass(obj)
|
|
|
|
|
|
def _lc_tool_call_to_fireworks_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_fireworks_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"],
|
|
},
|
|
}
|