app.py

import streamlit as st
import os
import re
import torch
import numpy as np
from pathlib import Path
import PyPDF2
from transformers import AutoTokenizer, AutoModelForCausalLM
from sentence_transformers import SentenceTransformer
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain_community.vectorstores import FAISS
from langchain.schema import Document
from langchain.embeddings import HuggingFaceEmbeddings

# Set page configuration
st.set_page_config(
    page_title="Vision 2030 Virtual Assistant",
    page_icon="🇸🇦",
    layout="wide"
)

# App title and description
st.title("Vision 2030 Virtual Assistant")
st.markdown("Ask questions about Saudi Vision 2030 goals, projects, and progress in Arabic or English.")

# Function definitions
@st.cache_resource
def load_model_and_tokenizer():
    """Load the ALLaM-7B model and tokenizer with error handling"""
    model_name = "ALLaM-AI/ALLaM-7B-Instruct-preview"
    st.info(f"Loading model: {model_name} (this may take a few minutes)")
    
    try:
        # First attempt with AutoTokenizer
        tokenizer = AutoTokenizer.from_pretrained(
            model_name,
            trust_remote_code=True,
            use_fast=False
        )
        
        # Load model with appropriate settings for ALLaM
        model = AutoModelForCausalLM.from_pretrained(
            model_name,
            torch_dtype=torch.bfloat16,
            trust_remote_code=True,
            device_map="auto",
        )
        
        st.success("Model loaded successfully!")
        
    except Exception as e:
        st.error(f"First loading attempt failed: {e}")
        st.info("Trying alternative loading approach...")
        
        # Try with specific tokenizer class if the first attempt fails
        from transformers import LlamaTokenizer
        
        tokenizer = LlamaTokenizer.from_pretrained(model_name)
        model = AutoModelForCausalLM.from_pretrained(
            model_name,
            torch_dtype=torch.float16,
            trust_remote_code=True,
            device_map="auto",
        )
        
        st.success("Model loaded successfully with LlamaTokenizer!")
    
    return model, tokenizer

def detect_language(text):
    """Detect if text is primarily Arabic or English"""
    arabic_chars = re.findall(r'[\u0600-\u06FF]', text)
    is_arabic = len(arabic_chars) > len(text) * 0.5
    return "arabic" if is_arabic else "english"

def process_pdfs():
    """Process uploaded PDF documents"""
    documents = []
    
    if 'uploaded_pdfs' in st.session_state and st.session_state.uploaded_pdfs:
        for pdf_file in st.session_state.uploaded_pdfs:
            try:
                # Save the uploaded file temporarily
                pdf_path = f"temp_{pdf_file.name}"
                with open(pdf_path, "wb") as f:
                    f.write(pdf_file.getbuffer())
                
                # Extract text
                text = ""
                with open(pdf_path, 'rb') as file:
                    reader = PyPDF2.PdfReader(file)
                    for page in reader.pages:
                        text += page.extract_text() + "\n\n"
                
                # Remove temporary file
                os.remove(pdf_path)
                
                if text.strip():  # If we got some text
                    doc = Document(
                        page_content=text,
                        metadata={"source": pdf_file.name, "filename": pdf_file.name}
                    )
                    documents.append(doc)
                    st.info(f"Successfully processed: {pdf_file.name}")
                else:
                    st.warning(f"No text extracted from {pdf_file.name}")
            except Exception as e:
                st.error(f"Error processing {pdf_file.name}: {e}")
    
    st.success(f"Processed {len(documents)} PDF documents")
    return documents

def create_vector_store(documents):
    """Split documents into chunks and create a FAISS vector store"""
    # Text splitter for breaking documents into chunks
    text_splitter = RecursiveCharacterTextSplitter(
        chunk_size=500,
        chunk_overlap=50,
        separators=["\n\n", "\n", ".", "!", "?", ",", " ", ""]
    )
    
    # Split documents into chunks
    chunks = []
    for doc in documents:
        doc_chunks = text_splitter.split_text(doc.page_content)
        # Preserve metadata for each chunk
        chunks.extend([
            Document(page_content=chunk, metadata=doc.metadata)
            for chunk in doc_chunks
        ])
    
    st.info(f"Created {len(chunks)} chunks from {len(documents)} documents")
    
    # Create a proper embedding function for LangChain
    embedding_function = HuggingFaceEmbeddings(
        model_name="sentence-transformers/paraphrase-multilingual-mpnet-base-v2"
    )
    
    # Create FAISS index
    vector_store = FAISS.from_documents(
        chunks,
        embedding_function
    )
    
    return vector_store

def retrieve_context(query, vector_store, top_k=5):
    """Retrieve most relevant document chunks for a given query"""
    # Search the vector store using similarity search
    results = vector_store.similarity_search_with_score(query, k=top_k)
    
    # Format the retrieved contexts
    contexts = []
    for doc, score in results:
        contexts.append({
            "content": doc.page_content,
            "source": doc.metadata.get("source", "Unknown"),
            "relevance_score": score
        })
    
    return contexts

def generate_response(query, contexts, model, tokenizer):
    """Generate a response using retrieved contexts with ALLaM-specific formatting"""
    # Auto-detect language
    language = detect_language(query)
    
    # Format the prompt based on language
    if language == "arabic":
        instruction = (
            "أنت مساعد افتراضي يهتم برؤية السعودية 2030. استخدم المعلومات التالية للإجابة على السؤال. "
            "إذا لم تعرف الإجابة، فقل بأمانة إنك لا تعرف."
        )
    else:  # english
        instruction = (
            "You are a virtual assistant for Saudi Vision 2030. Use the following information to answer the question. "
            "If you don't know the answer, honestly say you don't know."
        )
    
    # Combine retrieved contexts
    context_text = "\n\n".join([f"Document: {ctx['content']}" for ctx in contexts])
    
    # Format the prompt for ALLaM instruction format
    prompt = f"""<s>[INST] {instruction}

Context:
{context_text}

Question: {query} [/INST]</s>"""
    
    try:
        with st.spinner("Generating response..."):
            # Generate response with appropriate parameters for ALLaM
            inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
            
            # Generate with appropriate parameters
            outputs = model.generate(
                inputs.input_ids,
                attention_mask=inputs.attention_mask,
                max_new_tokens=512,
                temperature=0.7,
                top_p=0.9,
                do_sample=True,
                repetition_penalty=1.1
            )
            
            # Decode the response
            full_output = tokenizer.decode(outputs[0], skip_special_tokens=True)
            
            # Extract just the answer part (after the instruction)
            response = full_output.split("[/INST]")[-1].strip()
            
            # If response is empty for some reason, return the full output
            if not response:
                response = full_output
                
            return response, [ctx.get("source", "Unknown") for ctx in contexts]
        
    except Exception as e:
        st.error(f"Error during generation: {e}")
        # Fallback response
        return "I apologize, but I encountered an error while generating a response.", []

# Initialize the app state
if 'conversation_history' not in st.session_state:
    st.session_state.conversation_history = []

if 'vector_store' not in st.session_state:
    st.session_state.vector_store = None

if 'uploaded_pdfs' not in st.session_state:
    st.session_state.uploaded_pdfs = None

# PDF upload section
st.header("1. Upload Vision 2030 Documents")
uploaded_files = st.file_uploader("Upload PDF documents about Vision 2030", 
                                 type=["pdf"], 
                                 accept_multiple_files=True,
                                 help="Upload one or more PDF documents containing information about Vision 2030")

if uploaded_files:
    st.session_state.uploaded_pdfs = uploaded_files
    if st.button("Process PDFs"):
        documents = process_pdfs()
        if documents:
            with st.spinner("Creating vector database..."):
                st.session_state.vector_store = create_vector_store(documents)
            st.success("Vector database created successfully!")

# Load the model (cached)
model, tokenizer = load_model_and_tokenizer()

# Chat interface
st.header("2. Chat with the Vision 2030 Assistant")

# Display conversation history
for message in st.session_state.conversation_history:
    if message["role"] == "user":
        st.markdown(f"**You:** {message['content']}")
    else:
        st.markdown(f"**Assistant:** {message['content']}")
        if 'sources' in message and message['sources']:
            st.markdown(f"*Sources: {', '.join([os.path.basename(src) for src in message['sources']])}*")
    st.divider()

# Input for new question
user_input = st.text_input("Ask a question about Vision 2030 (in Arabic or English):", key="user_query")

# Examples
st.markdown("**Example questions:**")
examples_col1, examples_col2 = st.columns(2)
with examples_col1:
    st.markdown("- What is Saudi Vision 2030?")
    st.markdown("- What are the economic goals of Vision 2030?")
    st.markdown("- How does Vision 2030 support women's empowerment?")
with examples_col2:
    st.markdown("- ما هي رؤية السعودية 2030؟")
    st.markdown("- ما هي الأهداف الاقتصادية لرؤية 2030؟")
    st.markdown("- كيف تدعم رؤية 2030 تمكين المرأة السعودية؟")

# Process the user input
if user_input and st.session_state.vector_store:
    # Add user message to history
    st.session_state.conversation_history.append({"role": "user", "content": user_input})
    
    # Get response
    response, sources = generate_response(user_input, retrieve_context(user_input, st.session_state.vector_store), model, tokenizer)
    
    # Add assistant message to history
    st.session_state.conversation_history.append({"role": "assistant", "content": response, "sources": sources})
    
    # Rerun to update the UI
    st.experimental_rerun()

elif user_input and not st.session_state.vector_store:
    st.warning("Please upload and process Vision 2030 PDF documents first")

# Reset conversation button
if st.button("Reset Conversation") and len(st.session_state.conversation_history) > 0:
    st.session_state.conversation_history = []
    st.experimental_rerun()