add: main.py

main.py

@@ -5,11 +5,9 @@ from langchain_openai import ChatOpenAI, OpenAIEmbeddings
 from langchain_chroma import Chroma
 from langchain_text_splitters import RecursiveCharacterTextSplitter
 from langchain_core.documents import Document
-from langchain_tavily import TavilySearchResults
 from langchain.tools import tool
 from deepagents import create_deep_agent
 from deepagents.backends import FilesystemBackend, LocalShellBackend, CompositeBackend
-from langchain_core.messages import HumanMessage
 
 # Load environment variables
 load_dotenv()
@@ -30,12 +28,16 @@ embeddings = OpenAIEmbeddings(
 
 # ---------- Vector Store ----------
 CHROMA_DIR = "./chroma_db"
-vector_store = Chroma(collection_name="knowledge", embedding_function=embeddings, persist_directory=CHROMA_DIR)
+vector_store = Chroma(
+    collection_name="knowledge",
+    embedding_function=embeddings,
+    persist_directory=CHROMA_DIR,
+)
 
 # ---------- Document Loader ----------
 
 def load_documents(directory: str):
-    """Read .txt/.md files, split into chunks, and add to Chroma collection."""
+    """Read .txt/.md files, split into chunks and add to Chroma."""
     splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=200)
     docs = []
     for root, _, files in os.walk(directory):
@@ -54,28 +56,24 @@ def load_documents(directory: str):
 if not os.path.exists(CHROMA_DIR) or not os.listdir(CHROMA_DIR):
     load_documents("./documents")
 
-# ---------- Tavily Search ----------
-TAVILY_API_KEY = os.getenv("TAVILY_API_KEY")
-if not TAVILY_API_KEY:
-    raise ValueError("TAVILY_API_KEY not set in .env")
-
 # ---------- Tools ----------
 @tool
 def search_local_kb(query: str, top_k: int = 3) -> str:
-    """Semantic search in the local Chroma knowledge base."""
+    """Semantic search in the local knowledge base."""
     docs = vector_store.similarity_search(query, k=top_k)
     if not docs:
-        return "No relevant local knowledge found."
-    return "\n---\n".join([f"{d.metadata.get('source', 'unknown')}\n{d.page_content}" for d in docs])
+        return "No relevant information found in local knowledge base."
+    return "\n---\n".join(f"{d.metadata.get('source', 'unknown')}\n{d.page_content}" for d in docs)
 
 @tool
 def web_search(query: str) -> str:
     """Web search using Tavily."""
-    results = TavilySearchResults(api_key=TAVILY_API_KEY, max_results=3)
-    search_results = results.run(query)
-    if not search_results:
+    from langchain_tavily import TavilySearchResults
+    tavily = TavilySearchResults(max_results=3)
+    results = tavily.run(query)
+    if not results:
         return "No web results found."
-    return "\n---\n".join([f"{r['title']}\n{r['content']}" for r in search_results])
+    return "\n---\n".join(f"{r['title']}\n{r['content']}" for r in results)
 
 # ---------- Backend ----------
 backend = CompositeBackend([
@@ -88,17 +86,10 @@ agent = create_deep_agent(
     model=llm,
     tools=[search_local_kb, web_search],
     backend=backend,
-    system_prompt=(
-        "You are a RAG agent with a local knowledge base and web search capability. "
-        "When a user asks a question, decide whether the answer can be found in the local "
-        "knowledge base or requires up‑to‑date information from the web. Use the tool "
-        "search_local_kb for local queries and web_search for web queries. "
-        "Always indicate the source of the information in your final answer: "
-        "(chromadb) or (tavily)."
-    ),
+    system_prompt="You are a helpful assistant. For questions about local documents use the local knowledge base. For up‑to‑date facts use web search. Always state the source (chromadb or tavily) in your answer.",
 )
 
-# ---------- CLI ----------
+# ---------- Main Loop ----------
 async def main():
     print("RAG Agent ready. Type 'exit' to quit.")
     while True:
@@ -106,13 +97,14 @@ async def main():
         if user_input.lower() in {"exit", "quit"}:
             print("Goodbye!")
             break
+        # Invoke agent
         result = await agent.ainvoke(
             {"messages": [HumanMessage(content=user_input)]},
             {"configurable": {"thread_id": "session-1"}},
         )
-        # The agent returns a list of messages; the last is the assistant reply
-        reply = result["messages"][-1].content
-        print(f"\n{reply}")
+        # Extract last message content
+        answer = result["messages"][-1].content
+        print(f"\nОтвет:\n{answer}")
 
 if __name__ == "__main__":
     asyncio.run(main())