75 lines
2.2 KiB
Python
75 lines
2.2 KiB
Python
import os
|
|
from langchain_ollama import ChatOllama, OllamaEmbeddings
|
|
from langchain_qdrant import QdrantVectorStore
|
|
from langchain_text_splitters import RecursiveCharacterTextSplitter
|
|
from langchain_core.documents import Document
|
|
from langchain.tools import tool
|
|
from deepagents.backends import FilesystemBackend, LocalShellBackend, CompositeBackend
|
|
from deepagents import create_deep_agent
|
|
|
|
# Настройки окружения
|
|
OLLAMA_HOST = os.getenv("OLLAMA_HOST", "http://localhost:11434")
|
|
QDRANT_HOST = os.getenv("QDRANT_HOST", "http://localhost:6333")
|
|
|
|
# Эмбеддинги
|
|
embeddings = OllamaEmbeddings(
|
|
model="nomic-embed-text",
|
|
base_url=OLLAMA_HOST
|
|
)
|
|
|
|
# Векторная база
|
|
vector_store = QdrantVectorStore(
|
|
url=QDRANT_HOST,
|
|
collection_name="knowledge",
|
|
embedding_function=embeddings
|
|
)
|
|
|
|
# Чанкинг
|
|
chunker = RecursiveCharacterTextSplitter(
|
|
chunk_size=500,
|
|
chunk_overlap=50
|
|
)
|
|
|
|
@tool
|
|
def add_to_knowledge_base(content: str, title: str = "doc") -> str:
|
|
"""Add content to the knowledge base."""
|
|
chunks = chunker.split_text(content)
|
|
docs = [Document(page_content=chunk, metadata={"title": title}) for chunk in chunks]
|
|
vector_store.add_documents(docs)
|
|
return f"Added: {title} with {len(chunks)} chunks."
|
|
|
|
@tool
|
|
def search_knowledge_base(query: str, max_results: int = 3) -> str:
|
|
"""Search the knowledge base for relevant information."""
|
|
docs = vector_store.similarity_search(query, k=max_results)
|
|
if not docs:
|
|
return "No results."
|
|
return "\n".join(f"{i+1}. {doc.page_content}" for i, doc in enumerate(docs))
|
|
|
|
# LLM
|
|
llm = ChatOllama(
|
|
model="llama3",
|
|
base_url=OLLAMA_HOST
|
|
)
|
|
|
|
# Backend
|
|
backend = CompositeBackend([
|
|
LocalShellBackend(workspace_dir="./workspace"),
|
|
FilesystemBackend(),
|
|
])
|
|
|
|
# Системный промпт
|
|
system_prompt = (
|
|
"You are a helpful agent with access to a knowledge base. "
|
|
"Use the tools to search and add knowledge. "
|
|
"When you need to retrieve information, call search_knowledge_base. "
|
|
"When you need to store new information, call add_to_knowledge_base."
|
|
)
|
|
|
|
# Агент
|
|
agent = create_deep_agent(
|
|
model=llm,
|
|
tools=[add_to_knowledge_base, search_knowledge_base],
|
|
backend=backend,
|
|
system_prompt=system_prompt,
|
|
) |