import gradio as gr
import openai 
from getpass import getpass
import os
from langchain.vectorstores import FAISS
from langchain.embeddings import HuggingFaceEmbeddings

#OpenRouter API
openai.api_key = os.environ.get("OPENROUTER_API_KEY")
openai.api_base = "https://openrouter.ai/api/v1"

# Load vector DB
embedding_model = HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
db = FAISS.load_local("gita_vector_index", embedding_model, allow_dangerous_deserialization=True)
retriever = db.as_retriever()

# Chat function
def Talk2Gita(query, history):
    try:
        docs = retriever.get_relevant_documents(query)
        context = "\n\n".join([doc.page_content for doc in docs[:2]])

        response = openai.ChatCompletion.create(
            model="mistralai/mistral-7b-instruct",
            messages=[
                {"role": "system", "content": "You are a helpful assistant answering questions from the Bhagavad Gita."},
                {"role": "user", "content": f"""Context:
{context}

Question:
{query}"""}
            ],
            max_tokens=300,
        )

        return response["choices"][0]["message"]["content"]

    except Exception as e:
        return f" Error: {str(e)}"

# Launch Gradio app
gr.ChatInterface(Talk2Gita, title="Talk to the Bhagavad Gita").launch(share=True)