app.py

#!/usr/bin/env python3
"""
PDF Chat Web Application - Hugging Face Spaces Version
======================================================
A Streamlit web app for chatting with PDF documents using OpenAI.
Deployed on Hugging Face Spaces for public access.
"""

import streamlit as st
import os
import tempfile
import PyPDF2
from io import BytesIO
import requests
import json

# Page configuration
st.set_page_config(
    page_title="PDF Chat Assistant",
    page_icon="📄",
    layout="wide",
    initial_sidebar_state="expanded"
)

# Custom CSS for modern dark theme
st.markdown("""
<style>
    .stApp {
        background: linear-gradient(135deg, #1e1e1e 0%, #2d2d2d 100%);
        color: #ffffff;
    }
    
    .main-header {
        background: linear-gradient(135deg, #2d2d2d 0%, #3d3d3d 100%);
        padding: 2rem;
        border-radius: 15px;
        text-align: center;
        margin-bottom: 2rem;
        box-shadow: 0 4px 6px rgba(0,0,0,0.3);
        border: 1px solid #404040;
    }
    
    .feature-card {
        background: linear-gradient(135deg, #2d2d2d 0%, #3d3d3d 100%);
        padding: 1.5rem;
        border-radius: 10px;
        margin: 1rem 0;
        border: 1px solid #404040;
        box-shadow: 0 2px 4px rgba(0,0,0,0.2);
    }
    
    .user-message {
        background: linear-gradient(135deg, #007bff 0%, #0056b3 100%);
        color: white;
        padding: 1rem;
        border-radius: 15px;
        margin: 0.5rem 0;
        margin-left: 15%;
        box-shadow: 0 2px 4px rgba(0,123,255,0.3);
    }
    
    .ai-message {
        background: linear-gradient(135deg, #495057 0%, #343a40 100%);
        color: white;
        padding: 1rem;
        border-radius: 15px;
        margin: 0.5rem 0;
        margin-right: 15%;
        box-shadow: 0 2px 4px rgba(73,80,87,0.3);
    }
    
    .success-message {
        background: linear-gradient(135deg, #28a745 0%, #20c997 100%);
        color: white;
        padding: 1rem;
        border-radius: 10px;
        margin: 1rem 0;
        box-shadow: 0 2px 4px rgba(40,167,69,0.3);
    }
    
    .warning-message {
        background: linear-gradient(135deg, #ffc107 0%, #fd7e14 100%);
        color: #212529;
        padding: 1rem;
        border-radius: 10px;
        margin: 1rem 0;
        box-shadow: 0 2px 4px rgba(255,193,7,0.3);
    }
    
    .info-message {
        background: linear-gradient(135deg, #17a2b8 0%, #20c997 100%);
        color: white;
        padding: 1rem;
        border-radius: 10px;
        margin: 1rem 0;
        box-shadow: 0 2px 4px rgba(23,162,184,0.3);
    }
    
    .stTextInput > div > div > input {
        background-color: #3d3d3d;
        color: white;
        border: 2px solid #555;
        border-radius: 10px;
        padding: 0.5rem;
    }
    
    .stTextInput > div > div > input:focus {
        border-color: #007bff;
        box-shadow: 0 0 0 0.2rem rgba(0,123,255,0.25);
    }
    
    .stButton > button {
        background: linear-gradient(135deg, #007bff 0%, #0056b3 100%);
        color: white;
        border: none;
        border-radius: 10px;
        padding: 0.5rem 1rem;
        font-weight: 600;
        transition: all 0.3s ease;
    }
    
    .stButton > button:hover {
        background: linear-gradient(135deg, #0056b3 0%, #004085 100%);
        transform: translateY(-2px);
        box-shadow: 0 4px 8px rgba(0,123,255,0.3);
    }
    
    .upload-area {
        border: 2px dashed #007bff;
        border-radius: 15px;
        padding: 2rem;
        text-align: center;
        background: rgba(0,123,255,0.1);
        margin: 1rem 0;
    }
    
    .stats-container {
        display: grid;
        grid-template-columns: repeat(auto-fit, minmax(200px, 1fr));
        gap: 1rem;
        margin: 1rem 0;
    }
    
    .stat-card {
        background: linear-gradient(135deg, #343a40 0%, #495057 100%);
        padding: 1rem;
        border-radius: 10px;
        text-align: center;
        border: 1px solid #404040;
    }
</style>
""", unsafe_allow_html=True)

class PDFChatBot:
    def __init__(self):
        self.pdf_text = ""
        self.conversation_history = []
        self.pdf_pages = 0
        self.pdf_chars = 0
    
    def extract_pdf_text(self, pdf_file):
        """Extract text from PDF using PyPDF2"""
        try:
            pdf_reader = PyPDF2.PdfReader(pdf_file)
            text = ""
            page_count = len(pdf_reader.pages)
            
            for page in pdf_reader.pages:
                text += page.extract_text() + "\n"
            
            if not text.strip():
                return False, "Could not extract text from PDF. The PDF might contain only images or be password protected."
            
            self.pdf_text = text
            self.pdf_pages = page_count
            self.pdf_chars = len(text)
            
            return True, f"Successfully extracted text from {page_count} pages ({len(text):,} characters)!"
            
        except Exception as e:
            return False, f"Error reading PDF: {str(e)}"
    
    def ask_openai(self, question, api_key):
        """Ask OpenAI directly using the API"""
        try:
            # Limit context to prevent token limits
            context_limit = 3000
            context = f"Based on the following document content, please answer the question accurately and concisely.\n\nDocument:\n{self.pdf_text[:context_limit]}\n\nQuestion: {question}"
            
            headers = {
                "Authorization": f"Bearer {api_key}",
                "Content-Type": "application/json"
            }
            
            data = {
                "model": "gpt-3.5-turbo",
                "messages": [
                    {"role": "system", "content": "You are a helpful AI assistant that answers questions about documents. Be accurate, concise, and helpful. If you cannot find the answer in the document, say so clearly."},
                    {"role": "user", "content": context}
                ],
                "max_tokens": 1000,
                "temperature": 0.1
            }
            
            response = requests.post(
                "https://api.openai.com/v1/chat/completions",
                headers=headers,
                json=data,
                timeout=30
            )
            
            if response.status_code == 200:
                result = response.json()
                answer = result['choices'][0]['message']['content']
                
                # Store in conversation history
                self.conversation_history.append({"question": question, "answer": answer})
                
                return answer
            elif response.status_code == 401:
                return "❌ Invalid API key. Please check your OpenAI API key and try again."
            elif response.status_code == 429:
                return "⏳ Rate limit exceeded. Please wait a moment and try again."
            else:
                return f"❌ API Error: {response.status_code} - Please check your API key and try again."
                
        except requests.exceptions.Timeout:
            return "⏳ Request timed out. Please try again."
        except requests.exceptions.ConnectionError:
            return "🌐 Connection error. Please check your internet connection."
        except Exception as e:
            return f"❌ Error: {str(e)}"

def main():
    # Initialize session state
    if 'bot' not in st.session_state:
        st.session_state.bot = PDFChatBot()
    if 'messages' not in st.session_state:
        st.session_state.messages = []
    if 'pdf_processed' not in st.session_state:
        st.session_state.pdf_processed = False
    if 'uploaded_file_name' not in st.session_state:
        st.session_state.uploaded_file_name = ""
    
    # Header
    st.markdown("""
    <div class="main-header">
        <h1>📄 PDF Chat Assistant</h1>
        <p style="font-size: 1.2em; margin: 0.5rem 0;">Upload any PDF and start an intelligent conversation with your document!</p>
        <p style="color: #aaa; margin-top: 1rem;">Powered by OpenAI GPT • Built with ❤️ for Hugging Face Spaces</p>
    </div>
    """, unsafe_allow_html=True)
    
    # Sidebar
    with st.sidebar:
        st.markdown("### 🔑 Configuration")
        
        api_key = st.text_input(
            "OpenAI API Key", 
            type="password",
            help="Enter your OpenAI API key to start chatting with PDFs",
            placeholder="sk-..."
        )
        
        if api_key:
            st.success("✅ API Key Provided")
        else:
            st.error("❌ API Key Required")
        
        st.markdown("---")
        
        st.markdown("### 📚 How to Use")
        st.markdown("""
        1. **🔑 Enter your OpenAI API key** above
        2. **📤 Upload a PDF file** using the uploader
        3. **⏳ Wait** for text extraction (few seconds)
        4. **💬 Ask questions** about your document
        5. **🧠 Get AI-powered answers** instantly!
        """)
        
        st.markdown("---")
        
        st.markdown("### 🎯 Features")
        st.markdown("""
        • 📄 **PDF Text Extraction**
        • 🤖 **AI-Powered Q&A**
        • 💾 **Conversation Memory**
        • 🎨 **Beautiful Interface**
        • 🚀 **Fast & Responsive**
        • 🔒 **Privacy Focused**
        """)
        
        st.markdown("---")
        
        if st.button("🗑️ Clear Chat History", use_container_width=True):
            st.session_state.messages = []
            st.session_state.bot = PDFChatBot()
            st.session_state.pdf_processed = False
            st.session_state.uploaded_file_name = ""
            st.success("✅ Chat history cleared!")
            st.rerun()
        
        # API Key Info
        with st.expander("ℹ️ Get OpenAI API Key"):
            st.markdown("""
            **How to get your API key:**
            
            1. Go to [OpenAI Platform](https://platform.openai.com)
            2. Sign up or log in to your account
            3. Navigate to **API Keys** section
            4. Click **"Create new secret key"**
            5. Copy the key (starts with `sk-`)
            6. Paste it in the field above
            
            **Note:** Your API key is only used for this session and is not stored anywhere.
            """)
        
        # Status
        st.markdown("---")
        st.markdown("### 📊 Status")
        
        if st.session_state.pdf_processed:
            st.success("✅ PDF Ready")
            st.info(f"📄 {st.session_state.uploaded_file_name}")
            
            # Display PDF stats
            bot = st.session_state.bot
            st.markdown(f"""
            **📈 Document Stats:**
            - Pages: {bot.pdf_pages}
            - Characters: {bot.pdf_chars:,}
            - Conversations: {len(bot.conversation_history)}
            """)
        else:
            st.warning("⏳ No PDF loaded")
        
        if api_key:
            st.success("✅ API Connected")
        else:
            st.error("❌ API Key Missing")
    
    # Main content
    col1, col2 = st.columns([1, 2])
    
    with col1:
        st.markdown("### 📤 Upload Your PDF")
        
        uploaded_file = st.file_uploader(
            "Choose a PDF file",
            type="pdf",
            help="Upload any PDF document (max 200MB)",
            label_visibility="collapsed"
        )
        
        if not uploaded_file:
            st.markdown("""
            <div class="upload-area">
                <h3>📁 Drag & Drop Your PDF Here</h3>
                <p>Or click "Browse files" above to select a PDF</p>
                <br>
                <p><small>📋 Supported: PDF files up to 200MB<br>
                🔒 Your files are processed securely and not stored</small></p>
            </div>
            """, unsafe_allow_html=True)
        
        if uploaded_file and api_key:
            if not st.session_state.pdf_processed or st.session_state.uploaded_file_name != uploaded_file.name:
                with st.spinner("📖 Extracting text from your PDF..."):
                    success, message = st.session_state.bot.extract_pdf_text(uploaded_file)
                    
                    if success:
                        st.markdown(f"""
                        <div class="success-message">
                            <h4>✅ PDF Processed Successfully!</h4>
                            <p>{message}</p>
                        </div>
                        """, unsafe_allow_html=True)
                        
                        st.session_state.pdf_processed = True
                        st.session_state.uploaded_file_name = uploaded_file.name
                        
                        # Show file details
                        file_size = uploaded_file.size / 1024  # KB
                        bot = st.session_state.bot
                        
                        st.markdown(f"""
                        <div class="info-message">
                            <strong>📄 File:</strong> {uploaded_file.name}<br>
                            <strong>📊 Size:</strong> {file_size:.1f} KB<br>
                            <strong>📃 Pages:</strong> {bot.pdf_pages}<br>
                            <strong>📝 Characters:</strong> {bot.pdf_chars:,}<br>
                            <strong>🎯 Status:</strong> Ready for questions!
                        </div>
                        """, unsafe_allow_html=True)
                    else:
                        st.markdown(f"""
                        <div class="warning-message">
                            <h4>⚠️ Processing Failed</h4>
                            <p>{message}</p>
                        </div>
                        """, unsafe_allow_html=True)
        
        elif uploaded_file and not api_key:
            st.markdown("""
            <div class="warning-message">
                <h4>⚠️ API Key Required</h4>
                <p>Please enter your OpenAI API key in the sidebar to process the PDF.</p>
            </div>
            """, unsafe_allow_html=True)
        
        # Example questions
        if st.session_state.pdf_processed:
            st.markdown("### 💡 Try These Questions")
            
            example_questions = [
                "📋 What is this document about?",
                "📝 Summarize the main points",
                "🔍 What are the key details?",
                "📊 Give me important information",
                "❓ What questions can I ask?"
            ]
            
            for question in example_questions:
                if st.button(question, key=f"example_{question}", use_container_width=True):
                    # Trigger the question
                    question_text = question.split(" ", 1)[1]  # Remove emoji
                    st.session_state.pending_question = question_text
                    st.rerun()
    
    with col2:
        st.markdown("### 💬 Chat with Your PDF")
        
        # Chat container
        chat_container = st.container()
        
        with chat_container:
            if st.session_state.messages:
                for message in st.session_state.messages:
                    if message["role"] == "user":
                        st.markdown(f"""
                        <div class="user-message">
                            <strong>🧑 You:</strong> {message["content"]}
                        </div>
                        """, unsafe_allow_html=True)
                    else:
                        st.markdown(f"""
                        <div class="ai-message">
                            <strong>🤖 AI:</strong> {message["content"]}
                        </div>
                        """, unsafe_allow_html=True)
            else:
                if st.session_state.pdf_processed:
                    st.markdown("""
                    <div class="ai-message">
                        <strong>🤖 AI:</strong> Hello! I've analyzed your PDF document. What would you like to know about it? Feel free to ask any questions!
                    </div>
                    """, unsafe_allow_html=True)
                else:
                    st.markdown("""
                    <div class="feature-card" style="text-align: center; padding: 3rem;">
                        <h3>👋 Welcome to PDF Chat Assistant!</h3>
                        <p style="font-size: 1.1em; margin: 1rem 0;">Transform any PDF into an interactive conversation</p>
                        <br>
                        <div class="stats-container">
                            <div class="stat-card">
                                <h4>📄 Smart</h4>
                                <p>AI understands your documents</p>
                            </div>
                            <div class="stat-card">
                                <h4>⚡ Fast</h4>
                                <p>Instant answers to your questions</p>
                            </div>
                            <div class="stat-card">
                                <h4>🔒 Secure</h4>
                                <p>Your data stays private</p>
                            </div>
                        </div>
                        <br>
                        <p><strong>Get started:</strong> Add your API key and upload a PDF!</p>
                    </div>
                    """, unsafe_allow_html=True)
        
        # Input area
        st.markdown("---")
        
        # Check for pending question from example buttons
        if hasattr(st.session_state, 'pending_question'):
            user_question = st.session_state.pending_question
            del st.session_state.pending_question
        else:
            user_question = st.text_input(
                "Ask a question about your PDF:",
                placeholder="e.g., What are the main topics discussed in this document?",
                disabled=not st.session_state.pdf_processed,
                key="user_input"
            )
        
        col_btn1, col_btn2, col_btn3 = st.columns([2, 1, 1])
        
        with col_btn1:
            send_button = st.button("📤 Send Message", disabled=not st.session_state.pdf_processed, use_container_width=True)
        
        with col_btn2:
            if st.button("🔄 Refresh", disabled=not st.session_state.pdf_processed, use_container_width=True):
                st.rerun()
        
        # Process question
        if (send_button or hasattr(st.session_state, 'pending_question')) and user_question and st.session_state.pdf_processed and api_key:
            # Add user message
            st.session_state.messages.append({"role": "user", "content": user_question})
            
            # Get AI response
            with st.spinner("🤖 AI is analyzing your question..."):
                ai_response = st.session_state.bot.ask_openai(user_question, api_key)
            
            # Add AI response
            st.session_state.messages.append({"role": "assistant", "content": ai_response})
            
            st.rerun()
        
        elif send_button and not st.session_state.pdf_processed:
            st.warning("⚠️ Please upload and process a PDF first!")
        
        elif send_button and not api_key:
            st.warning("⚠️ Please enter your OpenAI API key in the sidebar!")

    # Footer
    st.markdown("---")
    st.markdown("""
    <div style="text-align: center; color: #888; padding: 2rem;">
        <h4>🚀 PDF Chat Assistant</h4>
        <p>Made with ❤️ using Streamlit • Powered by OpenAI GPT-3.5 • Hosted on 🤗 Hugging Face Spaces</p>
        <p><small>📄 Upload PDFs • 💬 Ask Questions • 🧠 Get AI Answers • 🔒 Privacy First</small></p>
        <br>
        <p><small>⭐ Like this app? Give it a star on Hugging Face Spaces!</small></p>
    </div>
    """, unsafe_allow_html=True)

if __name__ == "__main__":
    main()