import os, asyncio
from dotenv import load_dotenv
from langchain_openai import ChatOpenAI
from langchain_core.messages import HumanMessage
from langchain.tools import tool
from deepagents import create_deep_agent
from deepagents.backends import FilesystemBackend, LocalShellBackend, CompositeBackend
from langchain_chroma import Chroma
from langchain_text_splitters import RecursiveCharacterTextSplitter
from langchain_ollama import OllamaEmbeddings
from langchain_ollama import Ollama
from langchain_tavily import TavilySearchResults
from pathlib import Path

# Load env vars
load_dotenv()

# ---------- LLM ----------
llm = ChatOpenAI(
    model="openai/gpt-oss-20b:free",
    base_url="https://openrouter.ai/api/v1",
    api_key=os.getenv("OPENAI_API_KEY"),
    temperature=0.0,
)

# ---------- Vectorstore ----------
persist_dir = Path("./chroma_db")
persist_dir.mkdir(parents=True, exist_ok=True)

embeddings = OllamaEmbeddings(model="nomic-embed-text")
vectorstore = Chroma(
    collection_name="knowledge",
    embedding_function=embeddings,
    persist_directory=str(persist_dir),
)

# Load documents from ./documents
text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=200)
for file_path in Path("./documents").glob("**/*.*"):
    if file_path.suffix.lower() in {".txt", ".md"}:
        content = file_path.read_text(encoding="utf-8")
        docs = text_splitter.split_text(content)
        vectorstore.add_documents([{"page_content": d, "metadata": {"source": str(file_path)}} for d in docs])
vectorstore.persist()

# ---------- Tools ----------
@tool
def search_local_kb(query: str, top_k: int = 3) -> str:
    """Semantic search in local ChromaDB knowledge base."""
    docs = vectorstore.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])

@tool
def web_search(query: str) -> str:
    """Web search via Tavily."""
    tavily = TavilySearchResults(api_key=os.getenv("TAVILY_API_KEY"))
    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 results])

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

# ---------- Agent ----------
agent = create_deep_agent(
    model=llm,
    tools=[search_local_kb, web_search],
    backend=backend,
    system_prompt="You are a helpful RAG agent. For questions about local documents use search_local_kb, for up‑to‑date facts use web_search. Always state the source (chromadb or tavily) in your answer.",
)

async def main():
    print("Welcome to the RAG agent. Type 'exit' to quit.")
    while True:
        user_input = input("\nQuery: ")
        if user_input.lower() in {"exit", "quit"}:
            print("Goodbye!")
            break
        result = await agent.ainvoke(
            {"messages": [HumanMessage(content=user_input)]},
            {"configurable": {"thread_id": "session-1"}},
        )
        # The agent returns a list of messages; last is the assistant reply
        reply = result["messages"][-1].content
        print("\nAnswer:\n", reply)

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