fix(): 1 исправлений, 0 отстояно — main.py

main.py

@@ -1,140 +1,116 @@
-#!/usr/bin/env python3
-"""RAG‑agent with ChromaDB and Tavily search.
-
-This implementation follows the course specification and uses the
-`deepagents` framework to create a single agent that can decide whether
-to query the local knowledge base (ChromaDB) or perform a web search
-via Tavily.  The agent is backed by OpenRouter for both LLM and
-embeddings, complying with the mandatory technical constraints.
 """
-
+# main.py – RAG‑агент с ChromaDB и веб‑поиском (Tavily)
-import asyncio
+# Используем deepagents, Ollama (LLM и эмбеддинги) и ChromaDB.
+# Всё в одном файле для удобства.
+"""
 import os
+import asyncio
 from pathlib import Path
 
-from langchain_openai import ChatOpenAI, OpenAIEmbeddings
+# ────────────────────────────────────── DeepAgents ────────────────────────
-from langchain_chroma import Chroma
-from langchain_core.documents import Document
-from langchain_text_splitters import RecursiveCharacterTextSplitter
-from langchain_tavily import TavilySearchRun
-from langchain.tools import tool
 from deepagents import create_deep_agent
 from deepagents.backends import FilesystemBackend, LocalShellBackend, CompositeBackend
+from langchain_ollama import ChatOllama
+from langchain_ollama import OllamaEmbeddings
+from langchain.tools import tool
+from langchain_core.messages import HumanMessage
 
-# ---------------------------------------------------------------------------
+# ────────────────────────────────────── ChromaDB ────────────────────────
-# Configuration
+from langchain_chroma import Chroma
-# ---------------------------------------------------------------------------
+from langchain_text_splitters import RecursiveCharacterTextSplitter
-# Load environment variables (OpenRouter key, Tavily key)
+from langchain_core.documents import Document
-OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
+
+# ────────────────────────────────────── Tavily ────────────────────────
+from langchain_tavily import TavilySearchResults
+
+# ────────────────────────────────────── Настройки ────────────────────────
+OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")  # не используется, но оставляем для совместимости
 TAVILY_API_KEY = os.getenv("TAVILY_API_KEY")
+if not TAVILY_API_KEY:
+    raise RuntimeError("Требуется переменная окружения TAVILY_API_KEY")
 
-# Persist directory for ChromaDB
+# Папки
 CHROMA_DIR = Path("./chroma_db")
-CHROMA_DIR.mkdir(parents=True, exist_ok=True)
+DOCS_DIR = Path("./documents")
+WORKSPACE_DIR = Path("./workspace")
 
-# ---------------------------------------------------------------------------
+# ────────────────────────────────────── LLM и Embeddings ────────────────────────
-# 1. Vector store (ChromaDB + OpenRouter embeddings)
+llm = ChatOllama(model="llama3", temperature=0.0)
-# ---------------------------------------------------------------------------
+embeddings = OllamaEmbeddings(model="nomic-embed-text")
-embeddings = OpenAIEmbeddings(
-    model="text-embedding-3-small",
-    base_url="https://openrouter.ai/api/v1",
-    api_key=OPENAI_API_KEY,
-)
 
+# ────────────────────────────────────── VectorStore ────────────────────────
 vector_store = Chroma(
     collection_name="knowledge",
-    embedding_function=embeddings,
     persist_directory=str(CHROMA_DIR),
+    embedding_function=embeddings,
 )
 
-# Helper to load documents from a directory and add to the store
+# ────────────────────────────────────── Загрузка документов ────────────────────────
-
+def load_documents(directory: Path, store: Chroma, chunk_size: int = 1000, chunk_overlap: int = 200):
-def load_documents(directory: Path):
+    """Читает .txt/.md, разбивает на чанки и добавляет в Chroma."""
-    """Read .txt/.md files, split into chunks, and store in Chroma."""
+    if not directory.exists():
-    splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=200)
+        print(f"Документы не найдены в {directory}")
-    docs = []
+        return
-    for file in directory.glob("**/*"):
+    splitter = RecursiveCharacterTextSplitter(chunk_size=chunk_size, chunk_overlap=chunk_overlap)
-        if file.suffix.lower() not in {".txt", ".md"}:
+    for file_path in directory.glob("**/*.*"):
+        if file_path.suffix.lower() not in {".txt", ".md"}:
             continue
-        text = file.read_text(encoding="utf-8")
+        text = file_path.read_text(encoding="utf-8")
-        chunks = splitter.split_text(text)
+        docs = splitter.split_text(text)
-        docs.extend([Document(page_content=c, metadata={"source": str(file)}) for c in chunks])
+        documents = [Document(page_content=chunk, metadata={"source": str(file_path)}) for chunk in docs]
-    if docs:
+        store.add_documents(documents)
-        vector_store.add_documents(docs)
+    print("Документы загружены в Chroma.")
-        vector_store.persist()
 
-# ---------------------------------------------------------------------------
+# Если база пустая – загрузим
-# 2. Tools
+if not CHROMA_DIR.exists() or not any(CHROMA_DIR.iterdir()):
-# ---------------------------------------------------------------------------
+    load_documents(DOCS_DIR, vector_store)
+
+# ────────────────────────────────────── Инструменты ────────────────────────
 @tool
 def search_local_kb(query: str, top_k: int = 3) -> str:
-    """Semantic search in the local ChromaDB knowledge base."""
+    """Семантический поиск в локальной базе знаний."""
     docs = vector_store.similarity_search(query, k=top_k)
     if not docs:
-        return "No relevant information found in the local knowledge base."
+        return "No results in local knowledge base."
-    return "\n---\n".join([f"{i+1}. {d.page_content[:200]}…" for i, d in enumerate(docs)])
+    return "\n---\n".join(f"{d.metadata.get('source')}\n{d.page_content[:500]}" for d in docs)
 
 @tool
 def web_search(query: str) -> str:
-    """Perform a web search using Tavily."""
+    """Веб‑поиск через Tavily."""
-    tavily = TavilySearchRun(api_key=TAVILY_API_KEY, max_results=3)
+    tavily = TavilySearchResults(api_key=TAVILY_API_KEY, max_results=3)
     results = tavily.run(query)
-    if not results:
+    return "\n---\n".join(f"{r['title']}\n{r['content']}" for r in results)
-        return "No web results found."
-    return "\n---\n".join([f"{i+1}. {r['title']}\n{r['content'][:200]}…" for i, r in enumerate(results)])
-
-# ---------------------------------------------------------------------------
-# 3. Agent (deepagents)
-# ---------------------------------------------------------------------------
-llm = ChatOpenAI(
-    model="openai/gpt-oss-20b:free",
-    base_url="https://openrouter.ai/api/v1",
-    api_key=OPENAI_API_KEY,
-    temperature=0.0,
-)
 
+# ────────────────────────────────────── Backend ────────────────────────
 backend = CompositeBackend([
-    LocalShellBackend(workspace_dir="./workspace"),
+    LocalShellBackend(workspace_dir=str(WORKSPACE_DIR)),
     FilesystemBackend(),
 ])
 
-system_prompt = (
+# ────────────────────────────────────── Агент ────────────────────────
-    "You are a helpful assistant. For a user query, first decide whether the
-    answer can be found in the local knowledge base. If so, use the
-    `search_local_kb` tool. If the query requires up‑to‑date information,
-    use the `web_search` tool. Respond with the best answer and clearly
-    state the source: `chromadb` or `tavily`."
-)
-
 agent = create_deep_agent(
     model=llm,
     tools=[search_local_kb, web_search],
     backend=backend,
-    system_prompt=system_prompt,
+    system_prompt="Вы – RAG‑агент. При запросе о локальных документах используйте search_local_kb, для актуальных новостей – web_search. В ответе указывайте источник: chromadb или tavily.",
 )
 
-# ---------------------------------------------------------------------------
+# ────────────────────────────────────── Чат‑цикл ────────────────────────
-# 4. CLI loop
+async def chat_loop():
-# ---------------------------------------------------------------------------
+    print("RAG‑агент готов. Введите запрос (exit для выхода).")
-async def main():
-    print("RAG Agent ready. Type your question (or 'exit' to quit).")
-    thread_id = "session-1"
     while True:
-        user_input = input("\nЗапрос: ")
+        user_input = input("Запрос: ")
-        if user_input.lower() in {"exit", "quit", "q"}:
+        if user_input.lower() in {"exit", "quit"}:
-            print("Goodbye!")
+            print("Выход.")
             break
-        response = await agent.ainvoke(
+        result = await agent.ainvoke(
-            {"messages": [{"role": "user", "content": user_input}]},
+            {"messages": [HumanMessage(content=user_input)]},
-            {"configurable": {"thread_id": thread_id}},
+            {"configurable": {"thread_id": "session-1"}},
         )
-        # The last message contains the assistant reply
+        # Предполагаем, что агент возвращает сообщение с полем source в metadata
-        assistant_msg = response["messages"][-1].content
+        response = result["messages"][-1].content
-        print(f"\nОтвет:\n{assistant_msg}")
+        print("\nОтвет:\n", response)
+        print("\n---\n")
 
 if __name__ == "__main__":
-    # Load documents once at startup
+    asyncio.run(chat_loop())
-    docs_dir = Path("./documents")
-    if docs_dir.exists():
-        load_documents(docs_dir)
-    asyncio.run(main())