import os
import asyncio
import json
from pathlib import Path
from langchain_openai import ChatOpenAI, OpenAIEmbeddings
from langchain_chroma import Chroma
from langchain_core.documents import Document
from langchain.tools import tool
from deepagents import create_deep_agent
from deepagents.backends import FilesystemBackend, LocalShellBackend, CompositeBackend
from langchain_core.messages import HumanMessage

# DESIGN DECISION: Use ChromaDB for local vector store
# NECESSITY: The assignment explicitly requires ChromaDB + Ollama embeddings.
# OPTIMALITY: ChromaDB is lightweight, file-based, and integrates directly with LangChain.
# ALTERNATIVES CONSIDERED: QDrant would need a separate server process and more setup.

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

# Persist directory for Chroma
CHROMA_DIR = Path("./chroma_faq")

def load_faq_to_chroma():
    """
    Load .md files from data/ into ChromaDB.
    """
    vector_store = Chroma(
        collection_name="faq",
        embedding_function=embeddings,
        persist_directory=str(CHROMA_DIR),
    )
    if not CHROMA_DIR.exists() or not any(CHROMA_DIR.iterdir()):
        docs = []
        for md_file in Path("data").glob("*.md"):
            content = md_file.read_text(encoding="utf-8")
            docs.append(Document(page_content=content, metadata={"source": md_file.name}))
        vector_store.add_documents(docs)
        vector_store.persist()
    return vector_store

vector_store = load_faq_to_chroma()

@tool
def search_course_docs(query: str) -> str:
    """
    Search the knowledge base for relevant information.
    """
    docs = vector_store.similarity_search(query, k=3)
    return "\n\n".join(d.page_content for d in docs) if docs else "No results found."

@tool
def fetch_course_meta(query: str) -> str:
    """
    Fetch course metadata from a static JSON file.
    """
    meta_path = Path("meta.json")
    if not meta_path.exists():
        return "Metadata file not found."
    data = json.loads(meta_path.read_text(encoding="utf-8"))
    # Simple case-insensitive search in keys and values
    matches = []
    for key, value in data.items():
        if isinstance(value, dict):
            for subkey, subvalue in value.items():
                if query.lower() in subkey.lower() or query.lower() in str(subvalue).lower():
                    matches.append(f"{subkey}: {subvalue}")
        else:
            if query.lower() in key.lower() or query.lower() in str(value).lower():
                matches.append(f"{key}: {value}")
    return "\n".join(matches) if matches else "No metadata matches your query."

# LLM via OpenRouter
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,
)

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

system_prompt = (
    "You are a helpful FAQ bot. Use search_course_docs for questions about course materials. "
    "Use fetch_course_meta for questions about schedule or metadata. "
    "Do not use both tools unless necessary. "
    "Indicate source in your answer: source: chroma | mcp_meta."
)

agent = create_deep_agent(
    model=llm,
    tools=[search_course_docs, fetch_course_meta],
    backend=backend,
    system_prompt=system_prompt,
)

async def ask_agent(question: str, thread_id: str = "session-1") -> str:
    result = await agent.ainvoke(
        {"messages": [HumanMessage(content=question)]},
        {"configurable": {"thread_id": thread_id}},
    )
    return result["messages"][-1].content

async def main():
    preset_questions = [
        "What is covered in Lecture 1?",
        "Explain supervised learning.",
        "When is Lecture 2 scheduled?",
    ]
    print("=== Preset questions ===")
    for q in preset_questions:
        answer = await ask_agent(q)
        print(f"\nQ: {q}\nA: {answer}\n")
    print("=== Interactive mode (type 'exit' to quit) ===")
    while True:
        user_input = input("\nYour question: ")
        if user_input.lower() in {"exit", "quit"}:
            break
        answer = await ask_agent(user_input)
        print(f"\nAnswer: {answer}")

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