app.py

# ==============================================================================
# 1. IMPORTS & SETUP
# ==============================================================================
import os
import re
import difflib
import pandas as pd
from datetime import datetime
import gradio as gr
from transformers import pipeline, set_seed ,BitsAndBytesConfig

# --- Model Loading ---
# Read the Hugging Face token from the environment secrets
hf_token = os.environ.get("HUGGING_FACE_HUB_TOKEN")

# Check if the token is available
if not hf_token:
    print("Warning: HUGGING_FACE_HUB_TOKEN not found. This may fail if the model is private.")

print("Loading conversational model...")
# Use the token to authenticate when loading the model
bnb_config = BitsAndBytesConfig(
    load_in_4bit=True,               # Set to True for 4-bit quantization
    bnb_4bit_quant_type="nf4",       # "nf4" is usually the most stable
    bnb_4bit_use_double_quant=True,  # Double quantization helps reduce memory
    bnb_4bit_compute_dtype="float16" # Safer than float32, faster on GPUs
)

# Use the token to authenticate when loading the model
generator = pipeline(
    'text-generation',
    model='Warrior786/acc-llama-1b-merged-v10-lite',
    token=hf_token,
    device_map="auto",               # Let HF infer best device usage
    model_kwargs={"quantization_config": bnb_config}
)

set_seed(42)
print("Model loaded successfully.")



# --- File Paths ---
# Log conversations to a local CSV file within the Hugging Face Space
LOG_FILE = "ACC_chat_history.csv"


# ==============================================================================
# 2. CONSTANTS (Allowed Links & Emails)
# ==============================================================================

# List of all approved URLs for the chatbot to share
ALLOWED_LINKS = [
    "https://acceducate.org/", "https://acceducate.org/apply/", "https://acceducate.org/faqs/",
    "https://acceducate.org/programs/a-continuous-charity-acc-student-loan-refinance-program/",
    "https://acceducate.org/programs/", "https://acceducate.org/programs/help-a-student-with-sadaqah-jariyah/",
    "https://acceducate.org/programs/zakat/", "https://acceducate.org/programs/change-maker/",
    "https://acceducate.org/programs/break-the-chains-of-riba/", "https://acceducate.org/programs/women-empowerment/",
    "https://acceducate.org/programs/educational-justice/", "https://acceducate.org/ways-to-give/",
    "https://acceducate.org/ways-to-give/donate-stocks/", "https://acceducate.org/ways-to-give/open-an-educational-partnership-fund/",
    "https://acceducate.org/ways-to-give/legacy-fund/", "https://acceducate.org/ways-to-give/monthly-giving/",
    "https://acceducate.org/ways-to-give/support-accs-waqf-endowment-a-legacy-of-sustainable-charity/",
    "https://acceducate.org/ways-to-give/donate-by-zelle/", "https://acceducate.org/ways-to-give/matchinggift/",
    "https://acceducate.org/ways-to-give/donate-your-ira-to-acc/", "https://acceducate.org/ways-to-give/corporate-sponsored-volunteering/",
    "https://acceducate.org/ways-to-give/buy-coffee-give-sadaqah/", "https://acceducate.org/ways-to-give/host-an-iftar/",
    "https://acceducate.org/campaigns/grants/", "https://acceducate.org/student-initiated-fundraising-initiative/",
    "https://acceducate.org/about-us/", "https://acceducate.org/our-team/", "https://acceducate.org/financials/",
    "https://acceducate.org/careers/", "https://acceducate.org/why-acc/", "https://acceducate.org/invite-us/",
    "https://acceducate.org/events/", "https://acceducate.org/testimonials/", "https://www.facebook.com/acontinuouscharity/",
    "https://www.instagram.com/acontinuouscharity/", "https://www.youtube.com/channel/UCkq8wAvAqt54yzjzL_QIq3w",
    "https://twitter.com/accnational",
]

# Mapping of keywords to specific ACC pages for better matching
ACC_LINKS = {
    "About Us": "https://acceducate.org/about-us/", "Apply": "https://acceducate.org/apply/",
    "Careers": "https://acceducate.org/careers/", "Donate": "https://acceducate.org/ways-to-give/",
    "Events": "https://acceducate.org/events/", "FAQs": "https://acceducate.org/faqs/",
    "Zakat": "https://acceducate.org/programs/zakat/", "Loan Refinance": "https://acceducate.org/programs/a-continuous-charity-acc-student-loan-refinance-program/",
}

# Allowed email addresses
ALLOWED_EMAILS = ["donate@acceducate.org"]


# ==============================================================================
# 3. HELPER FUNCTIONS
# ==============================================================================

def normalize_acc_link(url: str, user_msg: str = "") -> str:
    """Cleans and restricts a given URL to the official list of allowed ACC links."""
    url = re.sub(r"https?://[^/\s]*acc[^/\s]*\.org", "https://acceducate.org", url)
    url = re.sub(r"(/(\w+))(\s*/\2)+", r"\1", url)
    if not url.endswith("/"): url += "/"
    if url in ALLOWED_LINKS: return url
    for keyword, link in ACC_LINKS.items():
        if keyword.lower() in user_msg.lower(): return link
    closest = difflib.get_close_matches(url, ALLOWED_LINKS, n=1, cutoff=0.6)
    return closest[0] if closest else "https://acceducate.org/"

def sanitize_response_links(text: str, user_msg: str = "") -> str:
    """Finds and normalizes all URLs in the bot's response."""
    text = re.sub(r"https?://\S+", lambda m: normalize_acc_link(m.group(0), user_msg), text)
    def replace_md(match):
        display, url = match.groups()
        final_url = normalize_acc_link(url, user_msg)
        norm_display = re.sub(r"^https?://", "", final_url).rstrip("/")
        return f"[{norm_display}]({final_url})"
    text = re.sub(r"\[([^\]]+)\]\((https?://[^\)]+)\)", replace_md, text)
    return text

def is_valid_email(email: str) -> bool:
    """Checks if a string is a validly formatted email address."""
    pattern = r"^[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,}$"
    return re.match(pattern, email) is not None

def log_to_csv(email, user_msg, bot_msg):
    """Saves the conversation turn to a CSV file."""
    timestamp = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
    new_row = pd.DataFrame([[timestamp, email, user_msg, bot_msg]], columns=["Time", "Email", "User", "Bot"])
    try:
        header = not os.path.exists(LOG_FILE)
        new_row.to_csv(LOG_FILE, mode='a', header=header, index=False)
    except Exception as e:
        print(f"Error logging to CSV: {e}")


# ==============================================================================
# 4. CORE CHATBOT LOGIC
# ==============================================================================

def chat_fn(message: str, history: list, user_email_state: str):
    """Main function to handle the chatbot's logic and conversation flow."""
    # --- Step 1: Handle initial state and email collection ---
    if user_email_state is None:
        if is_valid_email(message.strip()):
            user_email_state = message.strip()
            response = f"✅ Thank you! Your email `{user_email_state}` has been saved. How can I help you today?"
        else:
            response = "❌ That doesn't look like a valid email. Please enter a valid email address to continue."
        history.append((message, response))
        # Return empty textbox, updated history, and the new email state
        return "", history, user_email_state

    # --- Step 2: Generate a response using the AI model ---
    # Construct a conversation history string for the model
    # Use only the last 2 turns to keep the context concise
    history_context = history[-2:]
    conversation_text = ""
    for user_turn, bot_turn in history_context:
        if user_turn:
            conversation_text += f"User: {user_turn}\n"
        if bot_turn:
            conversation_text += f"Bot: {bot_turn}\n"
    
    prompt = f"{conversation_text}User: {message}\nBot:"

    # Generate response
    generated = generator(prompt, max_length=150, num_return_sequences=1, pad_token_id=generator.tokenizer.eos_token_id)
    raw_response = generated[0]['generated_text']

    # Extract only the last part of the response from the bot
    bot_response = raw_response.split("Bot:")[-1].strip()
    
    # Add a fallback if the model generates an empty response
    if not bot_response:
        bot_response = "I'm sorry, I'm not sure how to respond to that. Could you please rephrase?"

    # --- Step 3: Sanitize links and log the conversation ---
    sanitized_response = sanitize_response_links(bot_response, message)
    log_to_csv(user_email_state, message, sanitized_response)

    history.append((message, sanitized_response))
    return "", history, user_email_state

def welcome_message(history=None):
    """Initializes the chat with a welcome message."""
    if history is None:
        history = []
    history.append((None, " Welcome to the ACC Virtual Assistant! Please enter your email to start our chat."))
    return history



# ==============================================================================
# 5. GRADIO USER INTERFACE
# ==============================================================================

with gr.Blocks(
    css="""
    body, .gradio-container { background-color: #ffffff !important; color: #000000 !important; }
    #header { background-color: rgb(20, 90, 143); color: white; padding: 20px; border-radius: 12px; text-align: center; }
    #chatbot { height: 500px; }
    #footer { text-align: center; font-size: 14px; color: gray; margin-top: 15px; }
    """,
    title="IlmBot - ACC Assistant"
) as demo:
    # State object to store the user's email for the duration of the session
    user_email = gr.State(None)

    gr.Markdown(
        """
        <div id="header">
            <h1>📚 IlmBot — Your Knowledge Companion</h1>
            <p><b>A Continuous Charity (ACC)</b> Virtual Assistant</p>
        </div>
        """
    )

    chatbot = gr.Chatbot(label="💬 Conversation", elem_id="chatbot")
    msg_box = gr.Textbox(placeholder="Enter your email first, then your message...", label="✍️ Your Message")

    with gr.Row():
        send_btn = gr.Button("Send")
        clear_btn = gr.Button("Clear Chat")

    gr.Markdown(
        """
        <div id="footer">
            Powered by <b>ACC</b> • Learn more at <a href="https://acceducate.org/" target="_blank">acceducate.org</a>
        </div>
        """
    )

    # --- Event Handlers ---
    def user_submit(message, history, email_state):
        if not message.strip():
            return "", history, email_state
        return chat_fn(message, history, email_state)

    def clear_chat():
        """Resets the chat and clears the user's email state."""
        new_history = welcome_message([])
        return "", new_history, None # Clear textbox, history, and email state

    # Load the initial welcome message when the app starts
    demo.load(welcome_message, inputs=None, outputs=[chatbot])

    # Connect UI components to functions
    msg_box.submit(user_submit, [msg_box, chatbot, user_email], [msg_box, chatbot, user_email])
    send_btn.click(user_submit, [msg_box, chatbot, user_email], [msg_box, chatbot, user_email])
    clear_btn.click(clear_chat, None, [msg_box, chatbot, user_email])

# Launch the Gradio app
demo.launch(debug=True,share=True)