import os
from pdf_parser import extract_text_from_pdf
from rag import RAG
from llm import ask_llm
from utils import save_uploaded_file, cleanup_temp_files

rag = RAG()


def upload_document(file):
    if file is None:
        return "⚠️ Please select a valid PDF document.", "No document indexed."

    # 1. Save uploaded file to temp/
    temp_path = save_uploaded_file(file)
    if not temp_path or not os.path.exists(temp_path):
        return "❌ Error saving uploaded file.", "Upload failed."

    try:
        # 2. Extract text from PDF
        text = extract_text_from_pdf(temp_path)
        if not text or not text.strip():
            return "⚠️ No readable text found in PDF.", "Extraction empty."

        # 3. Create FAISS vector index
        rag.create_index(text)
        chunk_count = len(rag.chunks)

        filename = os.path.basename(temp_path)
        return (
            f"✅ **{filename}** indexed successfully ({chunk_count} text chunks ready).",
            f"📄 Active: **{filename}** ({chunk_count} chunks)"
        )
    finally:
        # 4. Clean up temporary folder
        cleanup_temp_files()


def chat(question, history):
    if history is None:
        history = []

    if not question or not question.strip():
        return history

    results = rag.search(question)

    answer = ask_llm(results, question, history=history)

    # Gradio 6 chatbot messages format
    history.append({"role": "user", "content": question})
    history.append({"role": "assistant", "content": answer})

    return history


def clear_chat():
    return []