import gradio as gr
import os
import google.generativeai as genai
from langchain_community.embeddings import HuggingFaceEmbeddings
from langchain_community.vectorstores import FAISS
from langchain_community.document_loaders import TextLoader
from langchain.text_splitter import RecursiveCharacterTextSplitter

# -------------------------------
# 1. Setup Gemini
# -------------------------------
genai.configure(api_key=os.getenv("GOOGLE_API_KEY"))
gemini_model = genai.GenerativeModel("gemini-2.5-flash")

# -------------------------------
# 2. Ensure about_me.txt exists
# -------------------------------
if not os.path.exists("about_me.txt"):
    with open("about_me.txt", "w") as f:
        f.write("""
        Hello! I am a portfolio chatbot. I can help answer questions about projects, skills, and experience.
        This is a sample portfolio text. Please replace this with your actual portfolio content.
        """)

# -------------------------------
# 3. Load data
# -------------------------------
try:
    loader = TextLoader("about_me.txt")
    docs = loader.load()
except Exception as e:
    print(f"Error loading document: {e}")
    from langchain.schema import Document
    docs = [Document(page_content="Hello! I am a portfolio chatbot ready to help you.")]

# -------------------------------
# 4. Split documents
# -------------------------------
text_splitter = RecursiveCharacterTextSplitter(
    chunk_size=500,
    chunk_overlap=50
)
split_docs = text_splitter.split_documents(docs)

# -------------------------------
# 5. Create embeddings & FAISS
# -------------------------------
print("Loading embeddings...")
embedding_model = HuggingFaceEmbeddings(
    model_name="sentence-transformers/all-MiniLM-L6-v2",
    model_kwargs={'device': 'cpu'}
)

print("Creating vector database...")
db = FAISS.from_documents(split_docs, embedding_model)

# -------------------------------
# 6. Ask function with Gemini refinement
# -------------------------------
def ask_bot_alternative(question: str):
    try:
        if not question.strip():
            return "Please ask me a question about the portfolio!"

        # Retrieve top documents
        retriever = db.as_retriever(search_kwargs={"k": 2})
        context_docs = retriever.get_relevant_documents(question)

        if not context_docs:
            return "I could not find an answer in the portfolio content."

        # Combine retrieved docs into context
        context = "\n".join([doc.page_content for doc in context_docs])

        # Send to Gemini for refinement
        prompt = f"""
        You are a helpful assistant. 
        Answer the following question using only the given context. 
        If the answer is not present, say "I don’t know".

        Question: {question}

        Context:
        {context}

        Final Answer:
        """

        response = gemini_model.generate_content(prompt)
        return response.text.strip()

    except Exception as e:
        return f"Sorry, I encountered an error: {str(e)[:200]}"

# -------------------------------
# 7. Gradio Interface
# -------------------------------
iface = gr.Interface(
    fn=ask_bot_alternative,
    inputs=gr.Textbox(
        label="Ask me about the portfolio",
        placeholder="What would you like to know?",
        max_lines=3
    ),
    outputs=gr.Textbox(
        label="Response",
        max_lines=10
    ),
    title="Portfolio Chatbot",
    description="Ask me questions about skills, projects, and experience!",
    examples=[
        "What are your technical skills?",
        "Tell me about your projects",
        "What is your background?"
    ],
    cache_examples=False,
    allow_flagging="never"
)

if __name__ == "__main__":
    print("Launching chatbot...")
    iface.launch(
        server_name="0.0.0.0",
        server_port=7860,
        show_error=True
    )