import gradio as gr
import faiss
import json
import numpy as np
from sentence_transformers import SentenceTransformer
from huggingface_hub import InferenceClient

"""
For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
"""
# Load Uber FAQ Data
with open("uber_faqs.json", "r") as f:
    faq_data = json.load(f)

faq_questions = [item["question"] for item in faq_data]
faq_answers = {item["question"]: item["answer"] for item in faq_data}

# Load Sentence Transformer Model for Embeddings
embedding_model = SentenceTransformer("sentence-transformers/all-MiniLM-L6-v2")
faq_embeddings = embedding_model.encode(faq_questions, convert_to_numpy=True)

# Create FAISS Index
index = faiss.IndexFlatL2(faq_embeddings.shape[1])
index.add(faq_embeddings)



def retrieve_uber_info(query):
    """Retrieve the most relevant Uber FAQ answer for the given query."""
    query_embedding = embedding_model.encode([query], convert_to_numpy=True)
    D, I = index.search(query_embedding, k=1)  # Get the closest match
    retrieved_question = faq_questions[I[0][0]]
    retrieved_answer = faq_answers[retrieved_question]
    return retrieved_answer


client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")


def respond(message, history, system_message, max_tokens, temperature, top_p):
    """Generate a response using Zephyr 7B while integrating retrieved Uber knowledge."""
    
    retrieved_answer = retrieve_uber_info(message)
    system_message += f"\n\nUber FAQ Context: {retrieved_answer}"

    messages = [{"role": "system", "content": system_message}]
    
    for val in history:
        if val[0]:
            messages.append({"role": "user", "content": val[0]})
        if val[1]:
            messages.append({"role": "assistant", "content": val[1]})
    
    messages.append({"role": "user", "content": message})

    response = ""
    for message in client.chat_completion(
        messages, max_tokens=max_tokens, stream=True, temperature=temperature, top_p=top_p
    ):
        token = message.choices[0].delta.content
        response += token
        yield response


"""
For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
"""
demo = gr.ChatInterface(
    respond,
    additional_inputs=[
        gr.Textbox(value="You are an Uber AI assistant. Only answer questions about Uber services, policies, pricing, and support. If a question is unrelated to Uber, say 'I can only help with Uber-related topics.'", label="System Instruction"),
        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
        gr.Slider(
            minimum=0.1,
            maximum=1.0,
            value=0.95,
            step=0.05,
            label="Top-p (nucleus sampling)",
        ),
    ],
)


if __name__ == "__main__":
    demo.launch()