docs/langchain/libs/partners/openai/langchain_openai/llms/base.py

from __future__ import annotations

import logging
import os
import sys
from typing import (
    AbstractSet,
    Any,
    AsyncIterator,
    Collection,
    Dict,
    Iterator,
    List,
    Literal,
    Mapping,
    Optional,
    Set,
    Tuple,
    Union,
)

import openai
import tiktoken
from langchain_core.callbacks import (
    AsyncCallbackManagerForLLMRun,
    CallbackManagerForLLMRun,
)
from langchain_core.language_models.llms import BaseLLM
from langchain_core.outputs import Generation, GenerationChunk, LLMResult
from langchain_core.pydantic_v1 import Field, SecretStr, root_validator
from langchain_core.utils import (
    convert_to_secret_str,
    get_from_dict_or_env,
    get_pydantic_field_names,
)
from langchain_core.utils.utils import build_extra_kwargs

logger = logging.getLogger(__name__)


def _update_token_usage(
    keys: Set[str], response: Dict[str, Any], token_usage: Dict[str, Any]
) -> None:
    """Update token usage."""
    _keys_to_use = keys.intersection(response["usage"])
    for _key in _keys_to_use:
        if _key not in token_usage:
            token_usage[_key] = response["usage"][_key]
        else:
            token_usage[_key] += response["usage"][_key]


def _stream_response_to_generation_chunk(
    stream_response: Dict[str, Any],
) -> GenerationChunk:
    """Convert a stream response to a generation chunk."""
    if not stream_response["choices"]:
        return GenerationChunk(text="")
    return GenerationChunk(
        text=stream_response["choices"][0]["text"],
        generation_info=dict(
            finish_reason=stream_response["choices"][0].get("finish_reason", None),
            logprobs=stream_response["choices"][0].get("logprobs", None),
        ),
    )


class BaseOpenAI(BaseLLM):
    """Base OpenAI large language model class."""

    client: Any = Field(default=None, exclude=True)  #: :meta private:
    async_client: Any = Field(default=None, exclude=True)  #: :meta private:
    model_name: str = Field(default="gpt-3.5-turbo-instruct", alias="model")
    """Model name to use."""
    temperature: float = 0.7
    """What sampling temperature to use."""
    max_tokens: int = 256
    """The maximum number of tokens to generate in the completion.
    -1 returns as many tokens as possible given the prompt and
    the models maximal context size."""
    top_p: float = 1
    """Total probability mass of tokens to consider at each step."""
    frequency_penalty: float = 0
    """Penalizes repeated tokens according to frequency."""
    presence_penalty: float = 0
    """Penalizes repeated tokens."""
    n: int = 1
    """How many completions to generate for each prompt."""
    best_of: int = 1
    """Generates best_of completions server-side and returns the "best"."""
    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(default=None, alias="api_key")
    """Automatically inferred from env var `OPENAI_API_KEY` if not provided."""
    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] = None
    batch_size: int = 20
    """Batch size to use when passing multiple documents to generate."""
    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."""
    logit_bias: Optional[Dict[str, float]] = Field(default_factory=dict)
    """Adjust the probability of specific tokens being generated."""
    max_retries: int = 2
    """Maximum number of retries to make when generating."""
    streaming: bool = False
    """Whether to stream the results or not."""
    allowed_special: Union[Literal["all"], AbstractSet[str]] = set()
    """Set of special tokens that are allowed。"""
    disallowed_special: Union[Literal["all"], Collection[str]] = "all"
    """Set of special tokens that are not allowed。"""
    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] = None
    """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] = None
    """Optional httpx.AsyncClient. Only used for async invocations. Must specify 
        http_client as well if you'd like a custom client for sync invocations."""

    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["streaming"] and values["n"] > 1:
            raise ValueError("Cannot stream results when n > 1.")
        if values["streaming"] and values["best_of"] > 1:
            raise ValueError("Cannot stream results when best_of > 1.")

        openai_api_key = get_from_dict_or_env(
            values, "openai_api_key", "OPENAI_API_KEY"
        )
        values["openai_api_key"] = (
            convert_to_secret_str(openai_api_key) if openai_api_key else None
        )
        values["openai_api_base"] = values["openai_api_base"] or os.getenv(
            "OPENAI_API_BASE"
        )
        values["openai_proxy"] = get_from_dict_or_env(
            values,
            "openai_proxy",
            "OPENAI_PROXY",
            default="",
        )
        values["openai_organization"] = (
            values["openai_organization"]
            or os.getenv("OPENAI_ORG_ID")
            or os.getenv("OPENAI_ORGANIZATION")
        )

        client_params = {
            "api_key": (
                values["openai_api_key"].get_secret_value()
                if values["openai_api_key"]
                else None
            ),
            "organization": values["openai_organization"],
            "base_url": values["openai_api_base"],
            "timeout": values["request_timeout"],
            "max_retries": values["max_retries"],
            "default_headers": values["default_headers"],
            "default_query": values["default_query"],
        }
        if not values.get("client"):
            sync_specific = {"http_client": values["http_client"]}
            values["client"] = openai.OpenAI(
                **client_params, **sync_specific
            ).completions
        if not values.get("async_client"):
            async_specific = {"http_client": values["http_async_client"]}
            values["async_client"] = openai.AsyncOpenAI(
                **client_params, **async_specific
            ).completions

        return values

    @property
    def _default_params(self) -> Dict[str, Any]:
        """Get the default parameters for calling OpenAI API."""
        normal_params: Dict[str, Any] = {
            "temperature": self.temperature,
            "top_p": self.top_p,
            "frequency_penalty": self.frequency_penalty,
            "presence_penalty": self.presence_penalty,
            "n": self.n,
            "logit_bias": self.logit_bias,
        }

        if self.max_tokens is not None:
            normal_params["max_tokens"] = self.max_tokens

        # Azure gpt-35-turbo doesn't support best_of
        # don't specify best_of if it is 1
        if self.best_of > 1:
            normal_params["best_of"] = self.best_of

        return {**normal_params, **self.model_kwargs}

    def _stream(
        self,
        prompt: str,
        stop: Optional[List[str]] = None,
        run_manager: Optional[CallbackManagerForLLMRun] = None,
        **kwargs: Any,
    ) -> Iterator[GenerationChunk]:
        params = {**self._invocation_params, **kwargs, "stream": True}
        self.get_sub_prompts(params, [prompt], stop)  # this mutates params
        for stream_resp in self.client.create(prompt=prompt, **params):
            if not isinstance(stream_resp, dict):
                stream_resp = stream_resp.model_dump()
            chunk = _stream_response_to_generation_chunk(stream_resp)

            if run_manager:
                run_manager.on_llm_new_token(
                    chunk.text,
                    chunk=chunk,
                    verbose=self.verbose,
                    logprobs=(
                        chunk.generation_info["logprobs"]
                        if chunk.generation_info
                        else None
                    ),
                )
            yield chunk

    async def _astream(
        self,
        prompt: str,
        stop: Optional[List[str]] = None,
        run_manager: Optional[AsyncCallbackManagerForLLMRun] = None,
        **kwargs: Any,
    ) -> AsyncIterator[GenerationChunk]:
        params = {**self._invocation_params, **kwargs, "stream": True}
        self.get_sub_prompts(params, [prompt], stop)  # this mutates params
        async for stream_resp in await self.async_client.create(
            prompt=prompt, **params
        ):
            if not isinstance(stream_resp, dict):
                stream_resp = stream_resp.model_dump()
            chunk = _stream_response_to_generation_chunk(stream_resp)

            if run_manager:
                await run_manager.on_llm_new_token(
                    chunk.text,
                    chunk=chunk,
                    verbose=self.verbose,
                    logprobs=(
                        chunk.generation_info["logprobs"]
                        if chunk.generation_info
                        else None
                    ),
                )
            yield chunk

    def _generate(
        self,
        prompts: List[str],
        stop: Optional[List[str]] = None,
        run_manager: Optional[CallbackManagerForLLMRun] = None,
        **kwargs: Any,
    ) -> LLMResult:
        """Call out to OpenAI's endpoint with k unique prompts.

        Args:
            prompts: The prompts to pass into the model.
            stop: Optional list of stop words to use when generating.

        Returns:
            The full LLM output.

        Example:
            .. code-block:: python

                response = openai.generate(["Tell me a joke."])
        """
        # TODO: write a unit test for this
        params = self._invocation_params
        params = {**params, **kwargs}
        sub_prompts = self.get_sub_prompts(params, prompts, stop)
        choices = []
        token_usage: Dict[str, int] = {}
        # Get the token usage from the response.
        # Includes prompt, completion, and total tokens used.
        _keys = {"completion_tokens", "prompt_tokens", "total_tokens"}
        system_fingerprint: Optional[str] = None
        for _prompts in sub_prompts:
            if self.streaming:
                if len(_prompts) > 1:
                    raise ValueError("Cannot stream results with multiple prompts.")

                generation: Optional[GenerationChunk] = None
                for chunk in self._stream(_prompts[0], stop, run_manager, **kwargs):
                    if generation is None:
                        generation = chunk
                    else:
                        generation += chunk
                assert generation is not None
                choices.append(
                    {
                        "text": generation.text,
                        "finish_reason": (
                            generation.generation_info.get("finish_reason")
                            if generation.generation_info
                            else None
                        ),
                        "logprobs": (
                            generation.generation_info.get("logprobs")
                            if generation.generation_info
                            else None
                        ),
                    }
                )
            else:
                response = self.client.create(prompt=_prompts, **params)
                if not isinstance(response, dict):
                    # V1 client returns the response in an PyDantic object instead of
                    # dict. For the transition period, we deep convert it to dict.
                    response = 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.get("error"):
                    raise ValueError(response.get("error"))

                choices.extend(response["choices"])
                _update_token_usage(_keys, response, token_usage)
                if not system_fingerprint:
                    system_fingerprint = response.get("system_fingerprint")
        return self.create_llm_result(
            choices,
            prompts,
            params,
            token_usage,
            system_fingerprint=system_fingerprint,
        )

    async def _agenerate(
        self,
        prompts: List[str],
        stop: Optional[List[str]] = None,
        run_manager: Optional[AsyncCallbackManagerForLLMRun] = None,
        **kwargs: Any,
    ) -> LLMResult:
        """Call out to OpenAI's endpoint async with k unique prompts."""
        params = self._invocation_params
        params = {**params, **kwargs}
        sub_prompts = self.get_sub_prompts(params, prompts, stop)
        choices = []
        token_usage: Dict[str, int] = {}
        # Get the token usage from the response.
        # Includes prompt, completion, and total tokens used.
        _keys = {"completion_tokens", "prompt_tokens", "total_tokens"}
        system_fingerprint: Optional[str] = None
        for _prompts in sub_prompts:
            if self.streaming:
                if len(_prompts) > 1:
                    raise ValueError("Cannot stream results with multiple prompts.")

                generation: Optional[GenerationChunk] = None
                async for chunk in self._astream(
                    _prompts[0], stop, run_manager, **kwargs
                ):
                    if generation is None:
                        generation = chunk
                    else:
                        generation += chunk
                assert generation is not None
                choices.append(
                    {
                        "text": generation.text,
                        "finish_reason": (
                            generation.generation_info.get("finish_reason")
                            if generation.generation_info
                            else None
                        ),
                        "logprobs": (
                            generation.generation_info.get("logprobs")
                            if generation.generation_info
                            else None
                        ),
                    }
                )
            else:
                response = await self.async_client.create(prompt=_prompts, **params)
                if not isinstance(response, dict):
                    response = response.model_dump()
                choices.extend(response["choices"])
                _update_token_usage(_keys, response, token_usage)
        return self.create_llm_result(
            choices,
            prompts,
            params,
            token_usage,
            system_fingerprint=system_fingerprint,
        )

    def get_sub_prompts(
        self,
        params: Dict[str, Any],
        prompts: List[str],
        stop: Optional[List[str]] = None,
    ) -> List[List[str]]:
        """Get the sub prompts for llm call."""
        if stop is not None:
            if "stop" in params:
                raise ValueError("`stop` found in both the input and default params.")
            params["stop"] = stop
        if params["max_tokens"] == -1:
            if len(prompts) != 1:
                raise ValueError(
                    "max_tokens set to -1 not supported for multiple inputs."
                )
            params["max_tokens"] = self.max_tokens_for_prompt(prompts[0])
        sub_prompts = [
            prompts[i : i + self.batch_size]
            for i in range(0, len(prompts), self.batch_size)
        ]
        return sub_prompts

    def create_llm_result(
        self,
        choices: Any,
        prompts: List[str],
        params: Dict[str, Any],
        token_usage: Dict[str, int],
        *,
        system_fingerprint: Optional[str] = None,
    ) -> LLMResult:
        """Create the LLMResult from the choices and prompts."""
        generations = []
        n = params.get("n", self.n)
        for i, _ in enumerate(prompts):
            sub_choices = choices[i * n : (i + 1) * n]
            generations.append(
                [
                    Generation(
                        text=choice["text"],
                        generation_info=dict(
                            finish_reason=choice.get("finish_reason"),
                            logprobs=choice.get("logprobs"),
                        ),
                    )
                    for choice in sub_choices
                ]
            )
        llm_output = {"token_usage": token_usage, "model_name": self.model_name}
        if system_fingerprint:
            llm_output["system_fingerprint"] = system_fingerprint
        return LLMResult(generations=generations, llm_output=llm_output)

    @property
    def _invocation_params(self) -> Dict[str, Any]:
        """Get the parameters used to invoke the model."""
        return self._default_params

    @property
    def _identifying_params(self) -> Mapping[str, Any]:
        """Get the identifying parameters."""
        return {**{"model_name": self.model_name}, **self._default_params}

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

    def get_token_ids(self, text: str) -> List[int]:
        """Get the token IDs using the tiktoken package."""
        if self.custom_get_token_ids is not None:
            return self.custom_get_token_ids(text)
        # tiktoken NOT supported for Python < 3.8
        if sys.version_info[1] < 8:
            return super().get_num_tokens(text)

        model_name = self.tiktoken_model_name or self.model_name
        try:
            enc = tiktoken.encoding_for_model(model_name)
        except KeyError:
            enc = tiktoken.get_encoding("cl100k_base")

        return enc.encode(
            text,
            allowed_special=self.allowed_special,
            disallowed_special=self.disallowed_special,
        )

    @staticmethod
    def modelname_to_contextsize(modelname: str) -> int:
        """Calculate the maximum number of tokens possible to generate for a model.

        Args:
            modelname: The modelname we want to know the context size for.

        Returns:
            The maximum context size

        Example:
            .. code-block:: python

                max_tokens = openai.modelname_to_contextsize("gpt-3.5-turbo-instruct")
        """
        model_token_mapping = {
            "gpt-4o": 128_000,
            "gpt-4o-2024-05-13": 128_000,
            "gpt-4": 8192,
            "gpt-4-0314": 8192,
            "gpt-4-0613": 8192,
            "gpt-4-32k": 32768,
            "gpt-4-32k-0314": 32768,
            "gpt-4-32k-0613": 32768,
            "gpt-3.5-turbo": 4096,
            "gpt-3.5-turbo-0301": 4096,
            "gpt-3.5-turbo-0613": 4096,
            "gpt-3.5-turbo-16k": 16385,
            "gpt-3.5-turbo-16k-0613": 16385,
            "gpt-3.5-turbo-instruct": 4096,
            "text-ada-001": 2049,
            "ada": 2049,
            "text-babbage-001": 2040,
            "babbage": 2049,
            "text-curie-001": 2049,
            "curie": 2049,
            "davinci": 2049,
            "text-davinci-003": 4097,
            "text-davinci-002": 4097,
            "code-davinci-002": 8001,
            "code-davinci-001": 8001,
            "code-cushman-002": 2048,
            "code-cushman-001": 2048,
        }

        # handling finetuned models
        if "ft-" in modelname:
            modelname = modelname.split(":")[0]

        context_size = model_token_mapping.get(modelname, None)

        if context_size is None:
            raise ValueError(
                f"Unknown model: {modelname}. Please provide a valid OpenAI model name."
                "Known models are: " + ", ".join(model_token_mapping.keys())
            )

        return context_size

    @property
    def max_context_size(self) -> int:
        """Get max context size for this model."""
        return self.modelname_to_contextsize(self.model_name)

    def max_tokens_for_prompt(self, prompt: str) -> int:
        """Calculate the maximum number of tokens possible to generate for a prompt.

        Args:
            prompt: The prompt to pass into the model.

        Returns:
            The maximum number of tokens to generate for a prompt.

        Example:
            .. code-block:: python

                max_tokens = openai.max_token_for_prompt("Tell me a joke.")
        """
        num_tokens = self.get_num_tokens(prompt)
        return self.max_context_size - num_tokens


class OpenAI(BaseOpenAI):
    """OpenAI large language models.

    To use, you should have the environment variable ``OPENAI_API_KEY``
    set with your API key, or pass it as a named parameter to the constructor.

    Any parameters that are valid to be passed to the openai.create call can be passed
    in, even if not explicitly saved on this class.

    Example:
        .. code-block:: python

            from langchain_openai import OpenAI

            model = OpenAI(model_name="gpt-3.5-turbo-instruct")
    """

    @classmethod
    def get_lc_namespace(cls) -> List[str]:
        """Get the namespace of the langchain object."""
        return ["langchain", "llms", "openai"]

    @classmethod
    def is_lc_serializable(cls) -> bool:
        """Return whether this model can be serialized by Langchain."""
        return True

    @property
    def _invocation_params(self) -> Dict[str, Any]:
        return {**{"model": self.model_name}, **super()._invocation_params}

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

    @property
    def lc_attributes(self) -> Dict[str, Any]:
        attributes: Dict[str, Any] = {}
        if self.openai_api_base:
            attributes["openai_api_base"] = self.openai_api_base

        if self.openai_organization:
            attributes["openai_organization"] = self.openai_organization

        if self.openai_proxy:
            attributes["openai_proxy"] = self.openai_proxy

        return attributes