fix: main.py — Агент с RAG-памятью
This commit is contained in:
@@ -1,21 +1,20 @@
|
|||||||
import os
|
import os
|
||||||
import asyncio
|
import asyncio
|
||||||
from pathlib import Path
|
|
||||||
from typing import List
|
|
||||||
|
|
||||||
from dotenv import load_dotenv
|
from dotenv import load_dotenv
|
||||||
from langchain_openai import ChatOpenAI, OpenAIEmbeddings
|
from langchain_openai import ChatOpenAI, OpenAIEmbeddings
|
||||||
from langchain_chroma import Chroma
|
from langchain_core.messages import HumanMessage
|
||||||
from langchain_core.documents import Document
|
from langchain_core.documents import Document
|
||||||
from langchain.text_splitter import RecursiveCharacterTextSplitter
|
|
||||||
from langchain.tools import tool
|
from langchain.tools import tool
|
||||||
from deepagents import create_deep_agent
|
from deepagents import create_deep_agent
|
||||||
from deepagents.backends import FilesystemBackend, LocalShellBackend, CompositeBackend
|
from deepagents.backends import FilesystemBackend, LocalShellBackend, CompositeBackend
|
||||||
from langchain_core.messages import HumanMessage
|
from langchain_qdrant import QdrantVectorStore
|
||||||
|
from langchain_text_splitters import RecursiveCharacterTextSplitter
|
||||||
|
from qdrant_client import QdrantClient
|
||||||
|
|
||||||
|
# Загрузка переменных окружения
|
||||||
load_dotenv()
|
load_dotenv()
|
||||||
|
|
||||||
# ---------- LLM ----------
|
# Инициализация LLM
|
||||||
llm = ChatOpenAI(
|
llm = ChatOpenAI(
|
||||||
model="openai/gpt-oss-20b:free",
|
model="openai/gpt-oss-20b:free",
|
||||||
base_url="https://openrouter.ai/api/v1",
|
base_url="https://openrouter.ai/api/v1",
|
||||||
@@ -23,147 +22,90 @@ llm = ChatOpenAI(
|
|||||||
temperature=0.0,
|
temperature=0.0,
|
||||||
)
|
)
|
||||||
|
|
||||||
# ---------- Vector Store ----------
|
# Инициализация эмбеддингов
|
||||||
embeddings = OpenAIEmbeddings(
|
embeddings = OpenAIEmbeddings(
|
||||||
model="text-embedding-3-small",
|
model="text-embedding-3-small",
|
||||||
base_url="https://openrouter.ai/api/v1",
|
base_url="https://openrouter.ai/api/v1",
|
||||||
api_key=os.getenv("OPENAI_API_KEY"),
|
api_key=os.getenv("OPENAI_API_KEY"),
|
||||||
)
|
)
|
||||||
|
|
||||||
vector_store = Chroma(
|
# Инициализация Qdrant
|
||||||
collection_name="knowledge",
|
client = QdrantClient(url="http://localhost:6333")
|
||||||
embedding_function=embeddings,
|
collection_name = "knowledge_base"
|
||||||
|
vector_store = QdrantVectorStore(
|
||||||
|
client=client,
|
||||||
|
collection_name=collection_name,
|
||||||
|
embeddings=embeddings,
|
||||||
)
|
)
|
||||||
|
|
||||||
# ---------- Text Splitter ----------
|
# Чанкинг
|
||||||
splitter = RecursiveCharacterTextSplitter(
|
splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=200)
|
||||||
chunk_size=1000,
|
|
||||||
chunk_overlap=200,
|
|
||||||
separators=["\n\n", "\n", " "],
|
|
||||||
)
|
|
||||||
|
|
||||||
# ---------- RAG Tools ----------
|
# Инструмент поиска
|
||||||
@tool
|
@tool
|
||||||
def search_knowledge_base(query: str, max_results: int = 3) -> str:
|
def search_knowledge_base(query: str, max_results: int) -> str:
|
||||||
"""
|
"""Semantic search in the knowledge base."""
|
||||||
Perform a semantic search in the knowledge base.
|
docs = vector_store.similarity_search(query, k=max_results)
|
||||||
Returns the concatenated contents of the most relevant documents.
|
|
||||||
"""
|
|
||||||
docs: List[Document] = vector_store.similarity_search(query, k=max_results)
|
|
||||||
if not docs:
|
if not docs:
|
||||||
return "No relevant documents found."
|
return "No results found."
|
||||||
return "\n---\n".join(doc.page_content for doc in docs)
|
return "\n".join(doc.page_content for doc in docs)
|
||||||
|
|
||||||
|
|
||||||
|
# Инструмент добавления
|
||||||
@tool
|
@tool
|
||||||
def add_to_knowledge_base(content: str, title: str = "document") -> str:
|
def add_to_knowledge_base(content: str, title: str) -> str:
|
||||||
"""
|
"""Add content to the knowledge base."""
|
||||||
Add a new document to the knowledge base.
|
|
||||||
The content will be split into chunks before indexing.
|
|
||||||
"""
|
|
||||||
chunks = splitter.split_text(content)
|
chunks = splitter.split_text(content)
|
||||||
docs = [
|
docs = [Document(page_content=chunk, metadata={"title": title}) for chunk in chunks]
|
||||||
Document(page_content=chunk, metadata={"title": title, "chunk_index": i})
|
|
||||||
for i, chunk in enumerate(chunks)
|
|
||||||
]
|
|
||||||
vector_store.add_documents(docs)
|
vector_store.add_documents(docs)
|
||||||
return f"Added {len(docs)} chunks from '{title}' to the knowledge base."
|
return f"Added {len(docs)} chunks for {title}."
|
||||||
|
|
||||||
|
# Backend для deepagents
|
||||||
# ---------- Backend ----------
|
backend = CompositeBackend([
|
||||||
backend = CompositeBackend(
|
|
||||||
[
|
|
||||||
LocalShellBackend(workspace_dir="./workspace"),
|
LocalShellBackend(workspace_dir="./workspace"),
|
||||||
FilesystemBackend(),
|
FilesystemBackend(),
|
||||||
]
|
])
|
||||||
)
|
|
||||||
|
|
||||||
# ---------- Agent ----------
|
# Создание агента
|
||||||
agent = create_deep_agent(
|
agent = create_deep_agent(
|
||||||
model=llm,
|
model=llm,
|
||||||
tools=[search_knowledge_base, add_to_knowledge_base],
|
tools=[search_knowledge_base, add_to_knowledge_base],
|
||||||
backend=backend,
|
backend=backend,
|
||||||
system_prompt=(
|
system_prompt="You are a helpful agent with access to a knowledge base. Use the tools to search and add information.",
|
||||||
"You are an AI assistant with access to a local knowledge base. "
|
|
||||||
"When you need factual information, use the provided tools: "
|
|
||||||
"`search_knowledge_base` to retrieve data and `add_to_knowledge_base` to store new documents. "
|
|
||||||
"Always cite sources from the knowledge base in your answers."
|
|
||||||
),
|
|
||||||
)
|
)
|
||||||
|
|
||||||
# ---------- Helper Functions ----------
|
async def main():
|
||||||
def load_documents_from_directory(directory: Path) -> None:
|
print("Interactive agent. Commands: /add, /search, /quit")
|
||||||
"""
|
|
||||||
Recursively read .txt files from the given directory and add them to the knowledge base.
|
|
||||||
"""
|
|
||||||
for file_path in directory.rglob("*.txt"):
|
|
||||||
try:
|
|
||||||
content = file_path.read_text(encoding="utf-8")
|
|
||||||
title = file_path.stem
|
|
||||||
add_to_knowledge_base(content, title)
|
|
||||||
print(f"Loaded {file_path}")
|
|
||||||
except Exception as e:
|
|
||||||
print(f"Failed to load {file_path}: {e}")
|
|
||||||
|
|
||||||
|
|
||||||
async def chat_loop() -> None:
|
|
||||||
"""
|
|
||||||
Simple CLI loop.
|
|
||||||
Commands:
|
|
||||||
/add <path> - add a text file or all txt files in a directory
|
|
||||||
/search <q> - search the knowledge base
|
|
||||||
/quit - exit
|
|
||||||
Anything else is sent to the agent as a user message.
|
|
||||||
"""
|
|
||||||
thread_id = "cli-session"
|
|
||||||
print("AI assistant ready. Type /quit to exit.")
|
|
||||||
while True:
|
while True:
|
||||||
user_input = input(">>> ").strip()
|
try:
|
||||||
|
user_input = input("> ").strip()
|
||||||
|
except EOFError:
|
||||||
|
break
|
||||||
if not user_input:
|
if not user_input:
|
||||||
continue
|
continue
|
||||||
if user_input.lower() == "/quit":
|
if user_input.startswith("/add"):
|
||||||
|
title = input("Title: ").strip()
|
||||||
|
content = input("Content: ").strip()
|
||||||
|
result = add_to_knowledge_base(content, title)
|
||||||
|
print(result)
|
||||||
|
elif user_input.startswith("/search"):
|
||||||
|
query = input("Query: ").strip()
|
||||||
|
max_str = input("Max results (int): ").strip()
|
||||||
|
try:
|
||||||
|
max_results = int(max_str)
|
||||||
|
except ValueError:
|
||||||
|
max_results = 3
|
||||||
|
result = search_knowledge_base(query, max_results)
|
||||||
|
print(result)
|
||||||
|
elif user_input.startswith("/quit"):
|
||||||
print("Goodbye!")
|
print("Goodbye!")
|
||||||
break
|
break
|
||||||
if user_input.startswith("/add"):
|
|
||||||
parts = user_input.split(maxsplit=1)
|
|
||||||
if len(parts) != 2:
|
|
||||||
print("Usage: /add <path>")
|
|
||||||
continue
|
|
||||||
path = Path(parts[1]).expanduser().resolve()
|
|
||||||
if path.is_dir():
|
|
||||||
load_documents_from_directory(path)
|
|
||||||
elif path.is_file() and path.suffix.lower() == ".txt":
|
|
||||||
content = path.read_text(encoding="utf-8")
|
|
||||||
add_to_knowledge_base(content, path.stem)
|
|
||||||
print(f"Added file {path}")
|
|
||||||
else:
|
else:
|
||||||
print("Provide a .txt file or a directory containing .txt files.")
|
# обычный диалог с агентом
|
||||||
continue
|
|
||||||
if user_input.startswith("/search"):
|
|
||||||
parts = user_input.split(maxsplit=1)
|
|
||||||
if len(parts) != 2:
|
|
||||||
print("Usage: /search <query>")
|
|
||||||
continue
|
|
||||||
query = parts[1]
|
|
||||||
result = search_knowledge_base(query)
|
|
||||||
print(f"Search results:\n{result}")
|
|
||||||
continue
|
|
||||||
|
|
||||||
# Normal conversation with the agent
|
|
||||||
try:
|
|
||||||
response = await agent.ainvoke(
|
response = await agent.ainvoke(
|
||||||
{"messages": [HumanMessage(content=user_input)]},
|
{"messages": [HumanMessage(content=user_input)]},
|
||||||
{"configurable": {"thread_id": thread_id}},
|
{"configurable": {"thread_id": "session-1"}},
|
||||||
)
|
)
|
||||||
answer = response["messages"][-1].content
|
print(response["messages"][-1].content)
|
||||||
print(answer)
|
|
||||||
except Exception as e:
|
|
||||||
print(f"Agent error: {e}")
|
|
||||||
|
|
||||||
|
|
||||||
if __name__ == "__main__":
|
if __name__ == "__main__":
|
||||||
# Optional: preload a default docs folder
|
asyncio.run(main())
|
||||||
default_dir = Path("./docs")
|
|
||||||
if default_dir.is_dir():
|
|
||||||
load_documents_from_directory(default_dir)
|
|
||||||
asyncio.run(chat_loop())
|
|
||||||
Reference in New Issue
Block a user