import os
import asyncio
from pathlib import Path
from dotenv import load_dotenv
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.tools import tool
from deepagents import create_deep_agent
from deepagents.backends import FilesystemBackend, LocalShellBackend, CompositeBackend

# Load environment variables
load_dotenv()

# ---------- LLM and Embeddings ----------
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,
)

embeddings = OpenAIEmbeddings(
    model="text-embedding-3-small",
    base_url="https://openrouter.ai/api/v1",
    api_key=os.getenv("OPENAI_API_KEY"),
)

# ---------- Vector Store ----------
CHROMA_DIR = Path("./chroma_db")
vector_store = Chroma(
    collection_name="knowledge",
    embedding_function=embeddings,
    persist_directory=str(CHROMA_DIR),
)

# ---------- Document Loader ----------

def load_documents(directory: str, vectorstore: Chroma):
    splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=200)
    docs = []
    for file_path in Path(directory).glob("**/*.*"):
        if file_path.suffix.lower() not in {".txt", ".md"}:
            continue
        text = file_path.read_text(encoding="utf-8")
        chunks = splitter.split_text(text)
        docs.extend([Document(page_content=c, metadata={"source": str(file_path)}) for c in chunks])
    if docs:
        vectorstore.add_documents(docs)
        vectorstore.persist()

# Load initial documents if collection is empty
if not CHROMA_DIR.exists() or not list(CHROMA_DIR.iterdir()):
    load_documents("documents", vector_store)

# ---------- Tools ----------
@tool
def search_local_kb(query: str, top_k: int = 3) -> str:
    """Semantic search in the local knowledge base."""
    docs = vector_store.similarity_search(query, k=top_k)
    if not 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."""
    from langchain_tavily import TavilySearchResults
    tavily = TavilySearchResults()
    results = tavily.run(query)
    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 assistant that can search both a local knowledge base and the web.\nWhen answering, always indicate the source: either 'chromadb' or 'tavily'.\nUse the appropriate tool based on the query context.",
)

# ---------- CLI ----------
async def main():
    print("RAG Agent with ChromaDB and Tavily. Type 'exit' to quit.")
    while True:
        user_input = input("\nЗапрос: ")
        if user_input.lower() in {"exit", "quit"}:
            print("Goodbye!")
            break
        result = await agent.ainvoke(
            {"messages": [{"role": "user", "content": user_input}]},
            {"configurable": {"thread_id": "session-1"}},
        )
        # Extract last message content
        content = result["messages"][-1]["content"]
        print(content)

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