import os
import asyncio
from dotenv import load_dotenv
from langchain_ollama import ChatOllama
from langchain_ollama import OllamaEmbeddings
from langchain_chroma import Chroma
from langchain_core.documents import Document
from langchain_text_splitters import RecursiveCharacterTextSplitter
from langchain.tools import tool
from deepagents import create_deep_agent
from deepagents.backends import FilesystemBackend, LocalShellBackend, CompositeBackend

# Load environment variables
load_dotenv()

# ---------- Vector Store ----------

def create_vectorstore(persist_directory="./chroma_db"):
    """Create a Chroma vector store with Ollama embeddings."""
    embeddings = OllamaEmbeddings(model="nomic-embed-text")
    return Chroma(persist_directory=persist_directory, embedding_function=embeddings)


def load_documents(directory, vectorstore):
    """Load .txt and .md files from *directory*, chunk them, and add to *vectorstore*.
    The function preserves the file name in metadata for later reference.
    """
    splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=200)
    for root, _, files in os.walk(directory):
        for fname in files:
            if fname.lower().endswith(('.txt', '.md')):
                path = os.path.join(root, fname)
                with open(path, 'r', encoding='utf-8') as f:
                    text = f.read()
                docs = splitter.split_text(text)
                documents = [Document(page_content=chunk, metadata={"source": fname}) for chunk in docs]
                vectorstore.add_documents(documents)

# ---------- Tools ----------

vectorstore = create_vectorstore()

@tool
def search_local_kb(query: str, top_k: int = 3) -> str:
    """Semantic search in the local knowledge base (ChromaDB)."""
    retriever = vectorstore.as_retriever(search_kwargs={"k": top_k})
    docs = retriever.get_relevant_documents(query)
    if not docs:
        return "No relevant local knowledge found."
    return "\n---\n".join([f"{doc.metadata.get('source', 'unknown')}\n{doc.page_content}" for doc in docs])

@tool
def web_search(query: str) -> str:
    """Web search using Tavily."""
    from tavily import TavilyClient
    client = TavilyClient(api_key=os.getenv("TAVILY_API_KEY"))
    results = client.search(query, max_results=3)
    if not results:
        return "No web results found."
    return "\n---\n".join([f"{r['title']}\n{r['content']}" for r in results])

# ---------- Agent ----------

llm = ChatOllama(model="llama3", temperature=0.0)

backend = CompositeBackend([
    LocalShellBackend(workspace_dir="./workspace"),
    FilesystemBackend(),
])

system_prompt = (
    "You are an AI assistant with access to two tools: "
    "search_local_kb for local knowledge and web_search for up-to-date information. "
    "When answering a user query, first decide which tool is appropriate. "
    "If the answer can be derived from the local documents, use search_local_kb; "
    "otherwise use web_search. "
    "Always indicate the source of the information in your response: "
    "[Local KB] or [Web Search]."
)

agent = create_deep_agent(
    model=llm,
    tools=[search_local_kb, web_search],
    backend=backend,
    system_prompt=system_prompt,
)

async def main():
    print("RAG Agent ready. Type 'exit' to quit.")
    while True:
        user_input = input("\nЗапрос: ")
        if user_input.strip().lower() == "exit":
            print("Goodbye!")
            break
        result = await agent.ainvoke(
            {"messages": [{"role": "user", "content": user_input}]},
            {"configurable": {"thread_id": "session-1"}},
        )
        # The last message is the assistant's reply
        reply = result["messages"][-1].content
        print(f"\n{reply}")

if __name__ == "__main__":
    asyncio.run(main())