from typing import (
TYPE_CHECKING,
Any,
AsyncIterator,
Callable,
Dict,
Iterator,
List,
Mapping,
Optional,
Tuple,
Union,
)
from pydantic import Field
from swarm.utils.logger import logger
######### helpers
if TYPE_CHECKING:
import tiktoken
def _import_tiktoken() -> Any:
try:
import tiktoken
except ImportError:
raise ValueError(
"Could not import tiktoken python package. "
"This is needed in order to calculate get_token_ids. "
"Please install it with `pip install tiktoken`."
)
return tiktoken
def _convert_message_to_dict(message: BaseMessage) -> dict:
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"]
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,
}
else:
raise ValueError(f"Got unknown type {message}")
if "name" in message.additional_kwargs:
message_dict["name"] = message.additional_kwargs["name"]
return message_dict
class BaseMessage:
"""Base message class."""
class HumanMessage(BaseMessage):
"""Human message class."""
def __init__(self, content: str):
self.role = "user"
self.content = content
class AIMessage(BaseMessage):
"""AI message class."""
def __init__(self, content: str, additional_kwargs: Optional[Dict[str, Any]] = None):
self.role = "assistant"
self.content = content
self.additional_kwargs = additional_kwargs or {}
class SystemMessage(BaseMessage):
"""System message class."""
def __init__(self, content: str):
self.role = "system"
self.content = content
class FunctionMessage(BaseMessage):
"""Function message class."""
def __init__(self, content: str, name: str):
self.role = "function"
self.content = content
self.name = name
class ChatMessage(BaseMessage):
"""Chat message class."""
def __init__(self, content: str, role: str):
self.role = role
self.content = content
class BaseMessageChunk:
"""Base message chunk class."""
class HumanMessageChunk(BaseMessageChunk):
"""Human message chunk class."""
def __init__(self, content: str):
self.role = "user"
self.content = content
class AIMessageChunk(BaseMessageChunk):
"""AI message chunk class."""
def __init__(self, content: str, additional_kwargs: Optional[Dict[str, Any]] = None):
self.role = "assistant"
self.content = content
self.additional_kwargs = additional_kwargs or {}
class SystemMessageChunk(BaseMessageChunk):
"""System message chunk class."""
def __init__(self, content: str):
self.role = "system"
self.content = content
class FunctionMessageChunk(BaseMessageChunk):
"""Function message chunk class."""
def __init__(self, content: str, name: str):
self.role = "function"
self.content = content
self.name = name
class ChatMessageChunk(BaseMessageChunk):
"""Chat message chunk class."""
def __init__(self, content: str, role: str):
self.role = role
self.content = content
def convert_openai_messages(messages: List[dict]) -> List[BaseMessage]:
"""Convert dictionaries representing OpenAI messages to LangChain format.
Args:
messages: List of dictionaries representing OpenAI messages
Returns:
List of LangChain BaseMessage objects.
"""
converted_messages = []
for m in messages:
role = m.get("role")
content = m.get("content", "")
if m.get("function_call"):
additional_kwargs = {"function_call": dict(m["function_call"])}
else:
additional_kwargs = {}
if role == "user":
converted_messages.append(HumanMessage(content=content))
elif role == "assistant":
converted_messages.append(AIMessage(content=content, additional_kwargs=additional_kwargs))
elif role == "system":
converted_messages.append(SystemMessage(content=content))
elif role == "function":
converted_messages.append(FunctionMessage(content=content, name=m["name"]))
else:
converted_messages.append(ChatMessage(content=content, role=role))
return converted_messages
class ChatGenerationChunk:
"""Chat generation chunk class."""
def __init__(self, message: BaseMessageChunk):
self.message = message
def __add__(self, other: "ChatGenerationChunk") -> "ChatGenerationChunk":
if isinstance(self.message, AIMessageChunk) and isinstance(other.message, AIMessageChunk):
combined_kwargs = {
**self.message.additional_kwargs,
**other.message.additional_kwargs,
}
return ChatGenerationChunk(
AIMessageChunk(content=self.message.content + other.message.content, additional_kwargs=combined_kwargs)
)
return ChatGenerationChunk(BaseMessageChunk(content=self.message.content + other.message.content))
@property
def content(self) -> str:
return self.message.content
@property
def additional_kwargs(self) -> Dict[str, Any]:
return getattr(self.message, "additional_kwargs", {})
class ChatResult:
"""Chat result class."""
def __init__(self, generations: List[ChatGenerationChunk]):
self.generations = generations
class BaseChatModel:
"""Base chat model class."""
def _convert_delta_to_message_chunk(self, _dict: Mapping[str, Any], default_class: type[BaseMessageChunk]) -> BaseMessageChunk:
role = _dict.get("role")
content = _dict.get("content") or ""
if _dict.get("function_call"):
additional_kwargs = {"function_call": dict(_dict["function_call"])}
else:
additional_kwargs = {}
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(self, _dict: Mapping[str, Any]) -> BaseMessage:
role = _dict["role"]
if role == "user":
return HumanMessage(content=_dict["content"])
elif role == "assistant":
content = _dict.get("content", "") or ""
if _dict.get("function_call"):
additional_kwargs = {"function_call": dict(_dict["function_call"])}
else:
additional_kwargs = {}
return AIMessage(content=content, additional_kwargs=additional_kwargs)
elif role == "system":
return SystemMessage(content=_dict["content"])
elif role == "function":
return FunctionMessage(content=_dict["content"], name=_dict["name"])
else:
return ChatMessage(content=_dict["content"], role=role)
def completion_with_retry(self, run_manager: Optional[Callable] = None, **kwargs: Any) -> Any:
"""Use tenacity to retry the completion call."""
retry_decorator = self._create_retry_decorator(self, run_manager=run_manager)
@retry_decorator
def _completion_with_retry(**kwargs: Any) -> Any:
return self.client.create(**kwargs)
return _completion_with_retry(**kwargs)
def _create_retry_decorator(
self,
llm: "ChatOpenAI",
run_manager: Optional[Callable] = None,
) -> Callable[[Any], Any]:
import openai
errors = [
openai.error.Timeout,
openai.error.APIError,
openai.error.APIConnectionError,
openai.error.RateLimitError,
openai.error.ServiceUnavailableError,
]
return create_base_retry_decorator(
error_types=errors, max_retries=llm.max_retries, run_manager=run_manager
)
def _create_message_dicts(
self, messages: List[BaseMessage], stop: Optional[List[str]]
) -> Tuple[List[Dict[str, Any]], Dict[str, Any]]:
params = self._client_params
if stop is not None:
if "stop" in params:
raise ValueError("`stop` found in both the input and default params.")
params["stop"] = stop
message_dicts = [_convert_message_to_dict(m) for m in messages]
return message_dicts, params
def _combine_llm_outputs(self, llm_outputs: List[Optional[dict]]) -> dict:
overall_token_usage: dict = {}
for output in llm_outputs:
if output is None:
# Happens in streaming
continue
token_usage = output["token_usage"]
for k, v in token_usage.items():
if k in overall_token_usage:
overall_token_usage[k] += v
else:
overall_token_usage[k] = v
return {"token_usage": overall_token_usage, "model_name": self.model_name}
class ChatOpenAI(BaseChatModel):
"""Wrapper around OpenAI Chat large language models.
To use, you should have the ``openai`` python package installed, and the
environment variable ``OPENAI_API_KEY`` set with your API key.
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.chat_models import ChatOpenAI
openai = ChatOpenAI(model_name="gpt-3.5-turbo")
"""
def __init__(
self,
model_name: str = "gpt-3.5-turbo",
temperature: float = 0.7,
model_kwargs: Optional[Dict[str, Any]] = None,
openai_api_key: Optional[str] = None,
openai_api_base: Optional[str] = None,
openai_organization: Optional[str] = None,
openai_proxy: Optional[str] = None,
request_timeout: Optional[Union[float, Tuple[float, float]]] = None,
max_retries: int = 6,
streaming: bool = False,
n: int = 1,
max_tokens: Optional[int] = None,
tiktoken_model_name: Optional[str] = None,
):
self.model_name = model_name
self.temperature = temperature
self.model_kwargs = model_kwargs or {}
self.openai_api_key = openai_api_key
self.openai_api_base = openai_api_base
self.openai_organization = openai_organization
self.openai_proxy = openai_proxy
self.request_timeout = request_timeout
self.max_retries = max_retries
self.streaming = streaming
self.n = n
self.max_tokens = max_tokens
self.tiktoken_model_name = tiktoken_model_name
@property
def lc_secrets(self) -> Dict[str, str]:
return {"openai_api_key": "OPENAI_API_KEY"}
@property
def lc_serializable(self) -> bool:
return True
@property
def _default_params(self) -> Dict[str, Any]:
"""Get the default parameters for calling OpenAI API."""
return {
"model": self.model_name,
"request_timeout": self.request_timeout,
"max_tokens": self.max_tokens,
"stream": self.streaming,
"n": self.n,
"temperature": self.temperature,
**self.model_kwargs,
}
@property
def _client_params(self) -> Dict[str, Any]:
"""Get the parameters used for the openai client."""
openai_creds: Dict[str, Any] = {
"api_key": self.openai_api_key,
"api_base": self.openai_api_base,
"organization": self.openai_organization,
"model": self.model_name,
}
if self.openai_proxy:
import openai
openai.proxy = {"http": self.openai_proxy, "https": self.openai_proxy}
return {**self._default_params, **openai_creds}
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 "openai-chat"
def _get_encoding_model(self) -> Tuple[str, tiktoken.Encoding]:
tiktoken_ = _import_tiktoken()
if self.tiktoken_model_name is not None:
model = self.tiktoken_model_name
else:
model = self.model_name
if model == "gpt-3.5-turbo":
# gpt-3.5-turbo may change over time.
# Returning num tokens assuming gpt-3.5-turbo-0301.
model = "gpt-3.5-turbo-0301"
elif model == "gpt-4":
# gpt-4 may change over time.
# Returning num tokens assuming gpt-4-0314.
model = "gpt-4-0314"
# Returns the number of tokens used by a list of messages.
try:
encoding = tiktoken_.encoding_for_model(model)
except KeyError:
logger.warning("Warning: model not found. Using cl100k_base encoding.")
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."""
# 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]) -> int:
"""Calculate num tokens for gpt-3.5-turbo and gpt-4 with tiktoken package.
Official documentation: https://github.com/openai/openai-cookbook/blob/
main/examples/How_to_format_inputs_to_ChatGPT_models.ipynb"""
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://github.com/openai/openai-python/blob/main/chatml.md 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():
# 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
def get_token_usage(self, text: str, prefix_tokens: Optional[List[str]] = None) -> int:
"""Get the number of tokens used by the provided text."""
token_ids = self.get_token_ids(text)
if prefix_tokens:
prefix_token_ids = self.get_token_ids(" ".join(prefix_tokens))
token_ids = prefix_token_ids + token_ids
return len(token_ids)
def completion_with_retry(
self, run_manager: Optional[Callable] = None, **kwargs: Any
) -> Any:
"""Use tenacity to retry the completion call."""
retry_decorator = self._create_retry_decorator(run_manager=run_manager)
@retry_decorator
def _completion_with_retry(**kwargs: Any) -> Any:
return self.client.create(**kwargs)
return _completion_with_retry(**kwargs)
async def acompletion_with_retry(
self, run_manager: Optional[Callable] = None, **kwargs: Any
) -> Any:
"""Use tenacity to retry the async completion call."""
retry_decorator = self._create_retry_decorator(run_manager=run_manager)
@retry_decorator
async def _completion_with_retry(**kwargs: Any) -> Any:
# Use OpenAI's async api https://github.com/openai/openai-python#async-api
return await self.client.acreate(**kwargs)
return await _completion_with_retry(**kwargs)
def _create_retry_decorator(
self,
run_manager: Optional[Callable] = None,
) -> Callable[[Any], Any]:
import openai
errors = [
openai.error.Timeout,
openai.error.APIError,
openai.error.APIConnectionError,
openai.error.RateLimitError,
openai.error.ServiceUnavailableError,
]
return create_base_retry_decorator(
error_types=errors, max_retries=self.max_retries, run_manager=run_manager
)
async def _astream(
self,
messages: List[BaseMessage],
stop: Optional[List[str]] = None,
run_manager: Optional[Callable] = None,
**kwargs: Any,
) -> AsyncIterator[ChatGenerationChunk]:
message_dicts, params = self._create_message_dicts(messages, stop)
params = {**params, **kwargs, "stream": True}
default_chunk_class = AIMessageChunk
async for chunk in await self.acompletion_with_retry(
messages=message_dicts, run_manager=run_manager, **params
):
if len(chunk["choices"]) == 0:
continue
delta = chunk["choices"][0]["delta"]
chunk = _convert_delta_to_message_chunk(delta, default_chunk_class)
default_chunk_class = chunk.__class__
#=================================
# from typing import (
# Any,
# Dict,
# List,
# Mapping,
# Optional,
# )
# import openai
# class ChatResult:
# """Wrapper for the result of the chat generation process."""
# def __init__(
# self,
# generations: List[ChatGeneration],
# llm_output: Optional[Mapping[str, Any]] = None,
# ):
# self.generations = generations
# self.llm_output = llm_output or {}
# class BaseMessage:
# """Base class for different types of messages."""
# def __init__(self, content: str):
# self.content = content
# class AIMessage(BaseMessage):
# """Message from the AI Assistant."""
# def __init__(self, content: str, additional_kwargs: Optional[Dict[str, Any]] = None):
# super().__init__(content)
# self.additional_kwargs = additional_kwargs or {}
# class HumanMessage(BaseMessage):
# """Message from the User."""
# pass
# class SystemMessage(BaseMessage):
# """System message."""
# pass
# class FunctionMessage(BaseMessage):
# """Function message."""
# def __init__(self, content: str, name: str):
# super().__init__(content)
# self.name = name
# class ChatGeneration:
# """Wrapper for the chat generation information."""
# def __init__(
# self, message: BaseMessage, generation_info: Optional[Mapping[str, Any]] = None
# ):
# self.message = message
# self.generation_info = generation_info or {}
# class ChatGenerationChunk:
# """Wrapper for a chunk of chat generation."""
# def __init__(self, message: BaseMessage):
# self.message = message
# def get_from_env_or_raise(var_name: str) -> str:
# value = os.getenv(var_name)
# if value is None:
# raise ValueError(f"Environment variable {var_name} is not set.")
# return value
# class OpenAI:
# """Wrapper around OpenAI Chat large language models.
# To use, you should have the ``openai`` python package installed, and the
# environment variable ``OPENAI_API_KEY`` set with your API key.
# Example:
# .. code-block:: python
# from langchain.chat_models import OpenAI
# openai = OpenAI(model_name="gpt-3.5-turbo")
# """
# def __init__(
# self,
# model_name: str = "gpt-3.5-turbo",
# temperature: float = 0.7,
# openai_api_key: Optional[str] = None,
# request_timeout: Optional[float] = None,
# max_retries: int = 6,
# ):
# self.model_name = model_name
# self.temperature = temperature
# self.openai_api_key = openai_api_key
# self.request_timeout = request_timeout
# self.max_retries = max_retries
# self._initialize_openai()
# def _initialize_openai(self):
# """Initialize the OpenAI client."""
# if self.openai_api_key is None:
# raise ValueError("OPENAI_API_KEY environment variable is not set.")
# openai.api_key = self.openai_api_key
# def _create_retry_decorator(self):
# """Create a decorator to handle API call retries."""
# errors = [
# openai.error.Timeout,
# openai.error.APIError,
# openai.error.APIConnectionError,
# openai.error.RateLimitError,
# openai.error.ServiceUnavailableError,
# ]
# def retry_decorator(func):
# @wraps(func)
# def wrapper(*args, **kwargs):
# for _ in range(self.max_retries):
# try:
# return func(*args, **kwargs)
# except tuple(errors):
# continue
# raise ValueError("Max retries reached. Unable to complete the API call.")
# return wrapper
# return retry_decorator
# def _create_message_dict(self, message: BaseMessage) -> Dict[str, Any]:
# """Convert a LangChain message to an OpenAI message dictionary."""
# role = message.role
# content = message.content
# message_dict = {"role": role, "content": content}
# if role == "assistant" and isinstance(message, AIMessage):
# message_dict["function_call"] = message.additional_kwargs.get("function_call", {})
# if role == "function" and isinstance(message, FunctionMessage):
# message_dict["name"] = message.name
# return message_dict
# def _create_message_dicts(self, messages: List[BaseMessage]) -> List[Dict[str, Any]]:
# """Convert a list of LangChain messages to a list of OpenAI message dictionaries."""
# return [self._create_message_dict(message) for message in messages]
# @retry_decorator
# def _openai_completion(self, messages: List[Dict[str, Any]], params: Dict[str, Any]) -> Any:
# """Call the OpenAI Chat Completion API."""
# response = openai.ChatCompletion.create(messages=messages, **params)
# return response
# def generate(
# self,
# messages: List[BaseMessage],
# stop: Optional[List[str]] = None,
# **kwargs: Any,
# ) -> ChatResult:
# """Generate a response using the OpenAI Chat model.
# Args:
# messages (List[BaseMessage]): List of messages in the conversation.
# stop (Optional[List[str]]): List of stop sequences to stop generation.
# Returns:
# ChatResult: The generated response wrapped in ChatResult object.
# """
# params = {
# "model": self.model_name,
# "temperature": self.temperature,
# "max_tokens": kwargs.get("max_tokens"),
# "stream": kwargs.get("streaming", False),
# "n": kwargs.get("n", 1),
# "request_timeout": kwargs.get("request_timeout", self.request_timeout),
# }
# messages_dicts = self._create_message_dicts(messages)
# response = self._openai_completion(messages_dicts, params)
# # Process the response and create ChatResult
# generations = []
# for choice in response["choices"]:
# message = self._convert_message(choice["message"])
# generation_info = {"finish_reason": choice.get("finish_reason")}
# generation = ChatGeneration(message=message, generation_info=generation_info)
# generations.append(generation)
# return ChatResult(generations=generations)
# def _convert_message(self, message_dict: Dict[str, Any]) -> BaseMessage:
# """Convert an OpenAI message dictionary to a LangChain message."""
# role = message_dict["role"]
# content = message_dict["content"]
# if role == "user":
# return HumanMessage(content=content)
# elif role == "assistant":
# additional_kwargs = message_dict.get("function_call", {})
# return AIMessage(content=content, additional_kwargs=additional_kwargs)
# elif role == "system":
# return SystemMessage(content=content)
# elif role == "function":
# name = message_dict.get("name", "")
# return FunctionMessage(content=content, name=name)
# else:
# raise ValueError(f"Invalid role found in the message: {role}")