You cannot select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
langchain/libs/community/langchain_community/chat_models/zhipuai.py

337 lines
11 KiB
Python

"""ZHIPU AI chat models wrapper."""
from __future__ import annotations
import asyncio
import json
import logging
from functools import partial
from typing import Any, Dict, Iterator, List, Optional, cast
from langchain_core.callbacks import CallbackManagerForLLMRun
from langchain_core.language_models.chat_models import (
BaseChatModel,
generate_from_stream,
)
from langchain_core.messages import AIMessage, AIMessageChunk, BaseMessage
from langchain_core.outputs import ChatGeneration, ChatGenerationChunk, ChatResult
from langchain_core.pydantic_v1 import BaseModel, Field
logger = logging.getLogger(__name__)
class ref(BaseModel):
"""Reference used in CharacterGLM."""
enable: bool = Field(True)
search_query: str = Field("")
class meta(BaseModel):
"""Metadata used in CharacterGLM."""
user_info: str = Field("")
bot_info: str = Field("")
bot_name: str = Field("")
user_name: str = Field("User")
class ChatZhipuAI(BaseChatModel):
"""
`ZHIPU AI` large language chat models API.
To use, you should have the ``zhipuai`` python package installed.
Example:
.. code-block:: python
from langchain_community.chat_models import ChatZhipuAI
zhipuai_chat = ChatZhipuAI(
temperature=0.5,
api_key="your-api-key",
model="chatglm_turbo",
)
"""
zhipuai: Any
zhipuai_api_key: Optional[str] = Field(default=None, alias="api_key")
"""Automatically inferred from env var `ZHIPUAI_API_KEY` if not provided."""
model: str = Field("chatglm_turbo")
"""
Model name to use.
-chatglm_turbo:
According to the input of natural language instructions to complete a
variety of language tasks, it is recommended to use SSE or asynchronous
call request interface.
-characterglm:
It supports human-based role-playing, ultra-long multi-round memory,
and thousands of character dialogues. It is widely used in anthropomorphic
dialogues or game scenes such as emotional accompaniments, game intelligent
NPCS, Internet celebrities/stars/movie and TV series IP clones, digital
people/virtual anchors, and text adventure games.
"""
temperature: float = Field(0.95)
"""
What sampling temperature to use. The value ranges from 0.0 to 1.0 and cannot
be equal to 0.
The larger the value, the more random and creative the output; The smaller
the value, the more stable or certain the output will be.
You are advised to adjust top_p or temperature parameters based on application
scenarios, but do not adjust the two parameters at the same time.
"""
top_p: float = Field(0.7)
"""
Another method of sampling temperature is called nuclear sampling. The value
ranges from 0.0 to 1.0 and cannot be equal to 0 or 1.
The model considers the results with top_p probability quality tokens.
For example, 0.1 means that the model decoder only considers tokens from the
top 10% probability of the candidate set.
You are advised to adjust top_p or temperature parameters based on application
scenarios, but do not adjust the two parameters at the same time.
"""
request_id: Optional[str] = Field(None)
"""
Parameter transmission by the client must ensure uniqueness; A unique
identifier used to distinguish each request, which is generated by default
by the platform when the client does not transmit it.
"""
streaming: bool = Field(False)
"""Whether to stream the results or not."""
incremental: bool = Field(True)
"""
When invoked by the SSE interface, it is used to control whether the content
is returned incremented or full each time.
If this parameter is not provided, the value is returned incremented by default.
"""
return_type: str = Field("json_string")
"""
This parameter is used to control the type of content returned each time.
- json_string Returns a standard JSON string.
- text Returns the original text content.
"""
ref: Optional[ref] = Field(None)
"""
This parameter is used to control the reference of external information
during the request.
Currently, this parameter is used to control whether to reference external
information.
If this field is empty or absent, the search and parameter passing format
is enabled by default.
{"enable": "true", "search_query": "history "}
"""
meta: Optional[meta] = Field(None)
"""Used in CharacterGLM"""
@property
def _identifying_params(self) -> Dict[str, Any]:
return {"model_name": self.model}
@property
def _llm_type(self) -> str:
"""Return the type of chat model."""
return "zhipuai"
@property
def lc_secrets(self) -> Dict[str, str]:
return {"zhipuai_api_key": "ZHIPUAI_API_KEY"}
@classmethod
def get_lc_namespace(cls) -> List[str]:
"""Get the namespace of the langchain object."""
return ["langchain", "chat_models", "zhipuai"]
@property
def lc_attributes(self) -> Dict[str, Any]:
attributes: Dict[str, Any] = {}
if self.model:
attributes["model"] = self.model
if self.streaming:
attributes["streaming"] = self.streaming
if self.return_type:
attributes["return_type"] = self.return_type
return attributes
def __init__(self, *args: Any, **kwargs: Any) -> None:
super().__init__(*args, **kwargs)
try:
import zhipuai
self.zhipuai = zhipuai
self.zhipuai.api_key = self.zhipuai_api_key
except ImportError:
raise RuntimeError(
"Could not import zhipuai package. "
"Please install it via 'pip install zhipuai'"
)
def invoke(self, prompt: Any) -> Any: # type: ignore[override]
if self.model == "chatglm_turbo":
return self.zhipuai.model_api.invoke(
model=self.model,
prompt=prompt,
top_p=self.top_p,
temperature=self.temperature,
request_id=self.request_id,
return_type=self.return_type,
)
elif self.model == "characterglm":
_meta = cast(meta, self.meta).dict()
return self.zhipuai.model_api.invoke(
model=self.model,
meta=_meta,
prompt=prompt,
request_id=self.request_id,
return_type=self.return_type,
)
return None
def sse_invoke(self, prompt: Any) -> Any:
if self.model == "chatglm_turbo":
return self.zhipuai.model_api.sse_invoke(
model=self.model,
prompt=prompt,
top_p=self.top_p,
temperature=self.temperature,
request_id=self.request_id,
return_type=self.return_type,
incremental=self.incremental,
)
elif self.model == "characterglm":
_meta = cast(meta, self.meta).dict()
return self.zhipuai.model_api.sse_invoke(
model=self.model,
prompt=prompt,
meta=_meta,
request_id=self.request_id,
return_type=self.return_type,
incremental=self.incremental,
)
return None
async def async_invoke(self, prompt: Any) -> Any:
loop = asyncio.get_running_loop()
partial_func = partial(
self.zhipuai.model_api.async_invoke, model=self.model, prompt=prompt
)
response = await loop.run_in_executor(
None,
partial_func,
)
return response
async def async_invoke_result(self, task_id: Any) -> Any:
loop = asyncio.get_running_loop()
response = await loop.run_in_executor(
None,
self.zhipuai.model_api.query_async_invoke_result,
task_id,
)
return response
def _generate(
self,
messages: List[BaseMessage],
stop: Optional[List[str]] = None,
run_manager: Optional[CallbackManagerForLLMRun] = None,
stream: Optional[bool] = None,
**kwargs: Any,
) -> ChatResult:
"""Generate a chat response."""
prompt: List = []
for message in messages:
if isinstance(message, AIMessage):
role = "assistant"
else: # For both HumanMessage and SystemMessage, role is 'user'
role = "user"
prompt.append({"role": role, "content": message.content})
should_stream = stream if stream is not None else self.streaming
if not should_stream:
response = self.invoke(prompt)
if response["code"] != 200:
raise RuntimeError(response)
content = response["data"]["choices"][0]["content"]
return ChatResult(
generations=[ChatGeneration(message=AIMessage(content=content))]
)
else:
stream_iter = self._stream(
prompt=prompt,
stop=stop,
run_manager=run_manager,
**kwargs,
)
return generate_from_stream(stream_iter)
async def _agenerate( # type: ignore[override]
self,
messages: List[BaseMessage],
stop: Optional[List[str]] = None,
run_manager: Optional[CallbackManagerForLLMRun] = None,
stream: Optional[bool] = False,
**kwargs: Any,
) -> ChatResult:
"""Asynchronously generate a chat response."""
prompt = []
for message in messages:
if isinstance(message, AIMessage):
role = "assistant"
else: # For both HumanMessage and SystemMessage, role is 'user'
role = "user"
prompt.append({"role": role, "content": message.content})
invoke_response = await self.async_invoke(prompt)
task_id = invoke_response["data"]["task_id"]
response = await self.async_invoke_result(task_id)
while response["data"]["task_status"] != "SUCCESS":
await asyncio.sleep(1)
response = await self.async_invoke_result(task_id)
content = response["data"]["choices"][0]["content"]
content = json.loads(content)
return ChatResult(
generations=[ChatGeneration(message=AIMessage(content=content))]
)
def _stream( # type: ignore[override]
self,
prompt: List[Dict[str, str]],
stop: Optional[List[str]] = None,
run_manager: Optional[CallbackManagerForLLMRun] = None,
**kwargs: Any,
) -> Iterator[ChatGenerationChunk]:
"""Stream the chat response in chunks."""
response = self.sse_invoke(prompt)
for r in response.events():
if r.event == "add":
delta = r.data
chunk = ChatGenerationChunk(message=AIMessageChunk(content=delta))
if run_manager:
run_manager.on_llm_new_token(delta, chunk=chunk)
yield chunk
elif r.event == "error":
raise ValueError(f"Error from ZhipuAI API response: {r.data}")