langchain/chat_models/fireworks.py

from typing import (
    Any,
    AsyncIterator,
    Callable,
    Dict,
    Iterator,
    List,
    Optional,
    Type,
    Union,
)

from langchain_core.messages import (
    AIMessage,
    AIMessageChunk,
    BaseMessage,
    BaseMessageChunk,
    ChatMessage,
    ChatMessageChunk,
    FunctionMessage,
    FunctionMessageChunk,
    HumanMessage,
    HumanMessageChunk,
    SystemMessage,
    SystemMessageChunk,
)
from langchain_core.outputs import ChatGeneration, ChatGenerationChunk, ChatResult
from langchain_core.pydantic_v1 import Field, SecretStr, root_validator
from langchain_core.utils import convert_to_secret_str

from langchain.adapters.openai import convert_message_to_dict
from langchain.callbacks.manager import (
    AsyncCallbackManagerForLLMRun,
    CallbackManagerForLLMRun,
)
from langchain.chat_models.base import BaseChatModel
from langchain.llms.base import create_base_retry_decorator
from langchain.utils.env import get_from_dict_or_env


def _convert_delta_to_message_chunk(
    _dict: Any, default_class: Type[BaseMessageChunk]
) -> BaseMessageChunk:
    """Convert a delta response to a message chunk."""
    role = _dict.role
    content = _dict.content or ""
    additional_kwargs: Dict = {}

    if role == "user" or default_class == HumanMessageChunk:
        return HumanMessageChunk(content=content)
    elif role == "assistant" or default_class == AIMessageChunk:
        return AIMessageChunk(content=content, additional_kwargs=additional_kwargs)
    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 or default_class == ChatMessageChunk:
        return ChatMessageChunk(content=content, role=role)
    else:
        return default_class(content=content)


def convert_dict_to_message(_dict: Any) -> BaseMessage:
    """Convert a dict response to a message."""
    role = _dict.role
    content = _dict.content or ""
    if role == "user":
        return HumanMessage(content=content)
    elif role == "assistant":
        content = _dict.content
        additional_kwargs: Dict = {}
        return AIMessage(content=content, additional_kwargs=additional_kwargs)
    elif role == "system":
        return SystemMessage(content=content)
    elif role == "function":
        return FunctionMessage(content=content, name=_dict.name)
    else:
        return ChatMessage(content=content, role=role)


class ChatFireworks(BaseChatModel):
    """Fireworks Chat models."""

    model: str = "accounts/fireworks/models/llama-v2-7b-chat"
    model_kwargs: dict = Field(
        default_factory=lambda: {
            "temperature": 0.7,
            "max_tokens": 512,
            "top_p": 1,
        }.copy()
    )
    fireworks_api_key: Optional[SecretStr] = None
    max_retries: int = 20
    use_retry: bool = True

    @property
    def lc_secrets(self) -> Dict[str, str]:
        return {"fireworks_api_key": "FIREWORKS_API_KEY"}

    @classmethod
    def is_lc_serializable(cls) -> bool:
        return True

    @root_validator()
    def validate_environment(cls, values: Dict) -> Dict:
        """Validate that api key in environment."""
        try:
            import fireworks.client
        except ImportError as e:
            raise ImportError(
                "Could not import fireworks-ai python package. "
                "Please install it with `pip install fireworks-ai`."
            ) from e
        fireworks_api_key = convert_to_secret_str(
            get_from_dict_or_env(values, "fireworks_api_key", "FIREWORKS_API_KEY")
        )
        fireworks.client.api_key = fireworks_api_key.get_secret_value()
        return values

    @property
    def _llm_type(self) -> str:
        """Return type of llm."""
        return "fireworks-chat"

    def _generate(
        self,
        messages: List[BaseMessage],
        stop: Optional[List[str]] = None,
        run_manager: Optional[CallbackManagerForLLMRun] = None,
        **kwargs: Any,
    ) -> ChatResult:
        message_dicts = self._create_message_dicts(messages)

        params = {
            "model": self.model,
            "messages": message_dicts,
            **self.model_kwargs,
        }
        response = completion_with_retry(
            self,
            self.use_retry,
            run_manager=run_manager,
            stop=stop,
            **params,
        )
        return self._create_chat_result(response)

    async def _agenerate(
        self,
        messages: List[BaseMessage],
        stop: Optional[List[str]] = None,
        run_manager: Optional[AsyncCallbackManagerForLLMRun] = None,
        **kwargs: Any,
    ) -> ChatResult:
        message_dicts = self._create_message_dicts(messages)
        params = {
            "model": self.model,
            "messages": message_dicts,
            **self.model_kwargs,
        }
        response = await acompletion_with_retry(
            self, self.use_retry, run_manager=run_manager, stop=stop, **params
        )
        return self._create_chat_result(response)

    def _combine_llm_outputs(self, llm_outputs: List[Optional[dict]]) -> dict:
        if llm_outputs[0] is None:
            return {}
        return llm_outputs[0]

    def _create_chat_result(self, response: Any) -> ChatResult:
        generations = []
        for res in response.choices:
            message = convert_dict_to_message(res.message)
            gen = ChatGeneration(
                message=message,
                generation_info=dict(finish_reason=res.finish_reason),
            )
            generations.append(gen)
        llm_output = {"model": self.model}
        return ChatResult(generations=generations, llm_output=llm_output)

    def _create_message_dicts(
        self, messages: List[BaseMessage]
    ) -> List[Dict[str, Any]]:
        message_dicts = [convert_message_to_dict(m) for m in messages]
        return message_dicts

    def _stream(
        self,
        messages: List[BaseMessage],
        stop: Optional[List[str]] = None,
        run_manager: Optional[CallbackManagerForLLMRun] = None,
        **kwargs: Any,
    ) -> Iterator[ChatGenerationChunk]:
        message_dicts = self._create_message_dicts(messages)
        default_chunk_class = AIMessageChunk
        params = {
            "model": self.model,
            "messages": message_dicts,
            "stream": True,
            **self.model_kwargs,
        }
        for chunk in completion_with_retry(
            self, self.use_retry, run_manager=run_manager, stop=stop, **params
        ):
            choice = chunk.choices[0]
            chunk = _convert_delta_to_message_chunk(choice.delta, default_chunk_class)
            finish_reason = choice.finish_reason
            generation_info = (
                dict(finish_reason=finish_reason) if finish_reason is not None else None
            )
            default_chunk_class = chunk.__class__
            chunk = ChatGenerationChunk(message=chunk, generation_info=generation_info)
            yield chunk
            if run_manager:
                run_manager.on_llm_new_token(chunk.text, chunk=chunk)

    async def _astream(
        self,
        messages: List[BaseMessage],
        stop: Optional[List[str]] = None,
        run_manager: Optional[AsyncCallbackManagerForLLMRun] = None,
        **kwargs: Any,
    ) -> AsyncIterator[ChatGenerationChunk]:
        message_dicts = self._create_message_dicts(messages)
        default_chunk_class = AIMessageChunk
        params = {
            "model": self.model,
            "messages": message_dicts,
            "stream": True,
            **self.model_kwargs,
        }
        async for chunk in await acompletion_with_retry_streaming(
            self, self.use_retry, run_manager=run_manager, stop=stop, **params
        ):
            choice = chunk.choices[0]
            chunk = _convert_delta_to_message_chunk(choice.delta, default_chunk_class)
            finish_reason = choice.finish_reason
            generation_info = (
                dict(finish_reason=finish_reason) if finish_reason is not None else None
            )
            default_chunk_class = chunk.__class__
            chunk = ChatGenerationChunk(message=chunk, generation_info=generation_info)
            yield chunk
            if run_manager:
                await run_manager.on_llm_new_token(token=chunk.text, chunk=chunk)


def conditional_decorator(
    condition: bool, decorator: Callable[[Any], Any]
) -> Callable[[Any], Any]:
    def actual_decorator(func: Callable[[Any], Any]) -> Callable[[Any], Any]:
        if condition:
            return decorator(func)
        return func

    return actual_decorator


def completion_with_retry(
    llm: ChatFireworks,
    use_retry: bool,
    *,
    run_manager: Optional[CallbackManagerForLLMRun] = None,
    **kwargs: Any,
) -> Any:
    """Use tenacity to retry the completion call."""
    import fireworks.client

    retry_decorator = _create_retry_decorator(llm, run_manager=run_manager)

    @conditional_decorator(use_retry, retry_decorator)
    def _completion_with_retry(**kwargs: Any) -> Any:
        return fireworks.client.ChatCompletion.create(
            **kwargs,
        )

    return _completion_with_retry(**kwargs)


async def acompletion_with_retry(
    llm: ChatFireworks,
    use_retry: bool,
    *,
    run_manager: Optional[AsyncCallbackManagerForLLMRun] = None,
    **kwargs: Any,
) -> Any:
    """Use tenacity to retry the async completion call."""
    import fireworks.client

    retry_decorator = _create_retry_decorator(llm, run_manager=run_manager)

    @conditional_decorator(use_retry, retry_decorator)
    async def _completion_with_retry(**kwargs: Any) -> Any:
        return await fireworks.client.ChatCompletion.acreate(
            **kwargs,
        )

    return await _completion_with_retry(**kwargs)


async def acompletion_with_retry_streaming(
    llm: ChatFireworks,
    use_retry: bool,
    *,
    run_manager: Optional[AsyncCallbackManagerForLLMRun] = None,
    **kwargs: Any,
) -> Any:
    """Use tenacity to retry the completion call for streaming."""
    import fireworks.client

    retry_decorator = _create_retry_decorator(llm, run_manager=run_manager)

    @conditional_decorator(use_retry, retry_decorator)
    async def _completion_with_retry(**kwargs: Any) -> Any:
        return fireworks.client.ChatCompletion.acreate(
            **kwargs,
        )

    return await _completion_with_retry(**kwargs)


def _create_retry_decorator(
    llm: ChatFireworks,
    run_manager: Optional[
        Union[AsyncCallbackManagerForLLMRun, CallbackManagerForLLMRun]
    ] = None,
) -> Callable[[Any], Any]:
    """Define retry mechanism."""
    import fireworks.client

    errors = [
        fireworks.client.error.RateLimitError,
        fireworks.client.error.InternalServerError,
        fireworks.client.error.BadGatewayError,
        fireworks.client.error.ServiceUnavailableError,
    ]
    return create_base_retry_decorator(
        error_types=errors, max_retries=llm.max_retries, run_manager=run_manager
    )