app.py

import gradio as gr
from huggingface_hub import InferenceClient
from datasets import load_dataset
import faiss
import numpy as np
import os
import time
import json 

# ✅ Ensure FAISS is installed
os.system("pip install faiss-cpu")

def log(message):
    print(f"✅ {message}")


import os
import json
from datasets import load_dataset

DATA_DIR = "data"
os.makedirs(DATA_DIR, exist_ok=True)  # Ensure directory exists

# ✅ List of datasets
datasets = {
    "sales": "goendalf666/sales-conversations",
    "blended": "blended_skill_talk",
    "dialog": "daily_dialog",
    "multiwoz": "multi_woz_v22",
}

# ✅ Save datasets to JSON
for name, hf_name in datasets.items():
    print(f"📥 Downloading {name} dataset...")
    dataset = load_dataset(hf_name)

    # Extract training data
    train_data = dataset["train"]

    # Convert dataset to list of dictionaries
    data_list = [dict(row) for row in train_data]

    # Save to JSON
    file_path = os.path.join(DATA_DIR, f"{name}.json")
    with open(file_path, "w") as f:
        json.dump(data_list, f, indent=2)

    print(f"✅ {name} dataset saved to {file_path}")



# ✅ Step 1: Run Embedding Script (Import and Run)
log("🚀 Running embeddings script...")
import embeddings  # This will automatically run embeddings.py

time.sleep(5)  # Wait for embeddings to be created

# ✅ Step 2: Check FAISS index
def check_faiss():
    index_path = "my_embeddings"  # Adjust if needed

    try:
        index = faiss.read_index(index_path)
        num_vectors = index.ntotal
        dim = index.d

        if num_vectors > 0:
            return f"📊 FAISS index contains {num_vectors} vectors.\n✅ Embedding dimension: {dim}"
        else:
            return "⚠️ No embeddings found in FAISS index!"

    except Exception as e:
        return f"❌ ERROR: Failed to load FAISS index - {e}"

log("🔍 Checking FAISS embeddings...")
faiss_status = check_faiss()
log(faiss_status)

# ✅ Step 3: Initialize chatbot
client = InferenceClient("mistralai/Mistral-7B-Instruct-v0.3")

def respond(message, history, system_message, max_tokens, temperature, top_p):
    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_completions(
        messages, max_tokens=max_tokens, stream=True, temperature=temperature, top_p=top_p
    ):
        token = message["choices"][0]["delta"]["content"]
        response += token
        yield response

# ✅ Step 4: Start Chatbot Interface
demo = gr.ChatInterface(
    respond,
    additional_inputs=[
        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
        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)"),
    ],
)

log("✅ All systems go! Launching chatbot...")
if __name__ == "__main__":
    demo.launch()