mirror of
https://github.com/hwchase17/langchain.git
synced 2025-11-03 17:54:10 +00:00
493e474063
- Move the API reference into the vercel build - Update api reference organization and styling
123 lines
3.7 KiB
Python
123 lines
3.7 KiB
Python
from typing import Dict, List, Tuple
|
|
|
|
from langchain.agents import (
|
|
AgentExecutor,
|
|
)
|
|
from langchain.agents.format_scratchpad import format_to_openai_functions
|
|
from langchain.agents.output_parsers import OpenAIFunctionsAgentOutputParser
|
|
from langchain_community.chat_models import ChatOpenAI
|
|
from langchain_community.embeddings import OpenAIEmbeddings
|
|
from langchain_community.tools.convert_to_openai import format_tool_to_openai_function
|
|
from langchain_community.tools.tavily_search import TavilySearchResults
|
|
from langchain_community.utilities.tavily_search import TavilySearchAPIWrapper
|
|
from langchain_community.vectorstores import FAISS
|
|
from langchain_core.documents import Document
|
|
from langchain_core.messages import AIMessage, HumanMessage
|
|
from langchain_core.prompts import (
|
|
ChatPromptTemplate,
|
|
MessagesPlaceholder,
|
|
)
|
|
from langchain_core.pydantic_v1 import BaseModel, Field
|
|
from langchain_core.runnables import Runnable, RunnableLambda, RunnableParallel
|
|
from langchain_core.tools import BaseTool, Tool
|
|
|
|
# Create the tools
|
|
search = TavilySearchAPIWrapper()
|
|
description = """"Useful for when you need to answer questions \
|
|
about current events or about recent information."""
|
|
tavily_tool = TavilySearchResults(api_wrapper=search, description=description)
|
|
|
|
|
|
def fake_func(inp: str) -> str:
|
|
return "foo"
|
|
|
|
|
|
fake_tools = [
|
|
Tool(
|
|
name=f"foo-{i}",
|
|
func=fake_func,
|
|
description=("a silly function that gets info " f"about the number {i}"),
|
|
)
|
|
for i in range(99)
|
|
]
|
|
ALL_TOOLS: List[BaseTool] = [tavily_tool] + fake_tools
|
|
|
|
# turn tools into documents for indexing
|
|
docs = [
|
|
Document(page_content=t.description, metadata={"index": i})
|
|
for i, t in enumerate(ALL_TOOLS)
|
|
]
|
|
|
|
vector_store = FAISS.from_documents(docs, OpenAIEmbeddings())
|
|
|
|
retriever = vector_store.as_retriever()
|
|
|
|
|
|
def get_tools(query: str) -> List[Tool]:
|
|
docs = retriever.invoke(query)
|
|
return [ALL_TOOLS[d.metadata["index"]] for d in docs]
|
|
|
|
|
|
assistant_system_message = """You are a helpful assistant. \
|
|
Use tools (only if necessary) to best answer the users questions."""
|
|
assistant_system_message = """You are a helpful assistant. \
|
|
Use tools (only if necessary) to best answer the users questions."""
|
|
prompt = ChatPromptTemplate.from_messages(
|
|
[
|
|
("system", assistant_system_message),
|
|
MessagesPlaceholder(variable_name="chat_history"),
|
|
("user", "{input}"),
|
|
MessagesPlaceholder(variable_name="agent_scratchpad"),
|
|
]
|
|
)
|
|
|
|
|
|
def llm_with_tools(input: Dict) -> Runnable:
|
|
return RunnableLambda(lambda x: x["input"]) | ChatOpenAI(temperature=0).bind(
|
|
functions=input["functions"]
|
|
)
|
|
|
|
|
|
def _format_chat_history(chat_history: List[Tuple[str, str]]):
|
|
buffer = []
|
|
for human, ai in chat_history:
|
|
buffer.append(HumanMessage(content=human))
|
|
buffer.append(AIMessage(content=ai))
|
|
return buffer
|
|
|
|
|
|
agent = (
|
|
RunnableParallel(
|
|
{
|
|
"input": lambda x: x["input"],
|
|
"chat_history": lambda x: _format_chat_history(x["chat_history"]),
|
|
"agent_scratchpad": lambda x: format_to_openai_functions(
|
|
x["intermediate_steps"]
|
|
),
|
|
"functions": lambda x: [
|
|
format_tool_to_openai_function(tool) for tool in get_tools(x["input"])
|
|
],
|
|
}
|
|
)
|
|
| {
|
|
"input": prompt,
|
|
"functions": lambda x: x["functions"],
|
|
}
|
|
| llm_with_tools
|
|
| OpenAIFunctionsAgentOutputParser()
|
|
)
|
|
|
|
# LLM chain consisting of the LLM and a prompt
|
|
|
|
|
|
class AgentInput(BaseModel):
|
|
input: str
|
|
chat_history: List[Tuple[str, str]] = Field(
|
|
..., extra={"widget": {"type": "chat", "input": "input", "output": "output"}}
|
|
)
|
|
|
|
|
|
agent_executor = AgentExecutor(agent=agent, tools=ALL_TOOLS).with_types(
|
|
input_type=AgentInput
|
|
)
|