app.py

import os
import sys
import gradio as gr
from dotenv import load_dotenv
import tempfile
import pandas as pd
import sqlite3
from langchain_core.prompts import ChatPromptTemplate
from langchain_groq import ChatGroq
import plotly.express as px
import time
import plotly.io as pio
import traceback
import base64
from io import BytesIO
import speech_recognition as sr
from gtts import gTTS
import re

# Load environment variables
load_dotenv()

# Add parent directory to path to import backend modules
sys.path.append(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))

from backend.main import DocumentAssistant

# Initialize the document assistant
document_assistant = DocumentAssistant()

# Initialize the LLM using the llama3-8b-8192 model from Groq
llm = ChatGroq(
    model="llama3-8b-8192",
    temperature=0,
    max_tokens=None,
    timeout=None,
    max_retries=2,
    verbose=True,
    api_key=os.getenv("GROQ_API_KEY")
)

# Database path for CSV data
DB_PATH = os.path.join(os.path.dirname(os.path.abspath(__file__)), "data", "csv_data.db")
os.makedirs(os.path.dirname(DB_PATH), exist_ok=True)

# Current context to track what we're working with
current_context = {
    "file_type": None,
    "file_name": None,
    "table_name": None
}

# Add a global variable to store the current plot
current_plot = None

# Define the prompt with examples for SQL query generation
query_prompt = ChatPromptTemplate.from_template("""
You are a SQL expert. Given a question about data in a table, write a SQLite-compatible SQL query to answer the question.

Important guidelines:
1. Use SQLite syntax (not PostgreSQL or MySQL)
2. For date functions, use strftime() instead of EXTRACT
   - Example: strftime('%Y', date_column) instead of EXTRACT(YEAR FROM date_column)
3. SQLite doesn't have TRUNCATE function, use CAST((column / bin_size) AS INT) * bin_size instead
4. For percentiles, use window functions or approximate methods
5. Keep queries efficient and focused on answering the specific question
6. Always use 'data_tab' as the table name

Question: {question}

SQL Query:
""")

# Define the prompt for interpreting the SQL query result
interpret_prompt = ChatPromptTemplate.from_messages(
    [
        ("system", "You are an experienced data analyst. Provide a concise, natural language answer based on the given data summary. If relevant, give key statistics, trends, or patterns."),
        ("human", "Question: {question}\nSQL Query: {sql_query}\nData Summary:\n{data_summary}")
    ]
)

# Add this after the query_prompt definition
visualization_prompt = ChatPromptTemplate.from_template("""
You are a data visualization expert. Given a question about visualizing data, write a SQLite-compatible SQL query that will retrieve the appropriate data for the visualization.

Important guidelines for SQLite syntax:
1. Use strftime() for date functions:
   - Year: strftime('%Y', date_column)
   - Month: strftime('%m', date_column)
   - Day: strftime('%d', date_column)
   - Hour: strftime('%H', date_column)

2. For histograms and binning:
   - Use: CAST((column / bin_size) AS INT) * bin_size
   - Example: CAST((trip_distance / 0.5) AS INT) * 0.5 AS distance_bin

3. For percentiles and statistics:
   - SQLite doesn't have built-in percentile functions
   - Use simple aggregations (MIN, MAX, AVG, COUNT) instead

4. For time series:
   - Group by date parts using strftime()
   - Example: strftime('%Y-%m-%d', pickup_datetime) AS day

5. Always use 'data_tab' as the table name

Question: {question}
Visualization type: {viz_type}

SQL Query:
""")

def process_text_query(query, history):
    """Process a text query and update chat history"""
    if not query:
        return "", history
    
    # Add the user's query to history
    history.append({"role": "user", "content": query})
    
    start_time = time.time()
    
    # Check if we're in CSV context
    if current_context["file_type"] == "csv" and current_context["table_name"]:
        try:
            # Connect to the database
            conn = sqlite3.connect(DB_PATH)
            
            # Get column information for context
            cursor = conn.cursor()
            cursor.execute(f"PRAGMA table_info({current_context['table_name']});")
            columns = [info[1] for info in cursor.fetchall()]
            columns_str = ", ".join(columns)
            
            # Create question with context
            question_with_context = f"The table 'data_tab' has columns: {columns_str}. {query}"
            
            # Generate SQL query using LLM
            ai_msg = query_prompt | llm
            sql_query = ai_msg.invoke({"question": question_with_context}).content.strip()
            
            print(f"Generated SQL Query: {sql_query}")
            
            # Check if this is a visualization request
            is_visualization = any(word in query.lower() for word in ['plot', 'graph', 'chart', 'visualize', 'visualization', 'trend'])
            
            try:
                # Execute the query
                result_df = pd.read_sql_query(sql_query, conn)
                
                # Generate data summary
                if not result_df.empty:
                    data_summary = result_df.describe(include='all').to_string()
                    
                    # For small result sets, include the actual data
                    if len(result_df) <= 10:
                        data_summary += f"\n\nFull Results:\n{result_df.to_string()}"
                    else:
                        data_summary += f"\n\nFirst 5 rows:\n{result_df.head(5).to_string()}"
                else:
                    data_summary = "No relevant data found."
                
                # Generate interpretation
                answer_chain = interpret_prompt | llm
                interpretation = answer_chain.invoke({
                    "question": query,
                    "sql_query": sql_query,
                    "data_summary": data_summary
                }).content.strip()
                
                # Create the response
                response = f"**SQL Query:**\n```sql\n{sql_query}\n```\n\n"
                
                if not result_df.empty:
                    if len(result_df) > 10:
                        response += f"**Results (first 5 of {len(result_df)} rows):**\n```\n{result_df.head(5).to_string()}\n```\n\n"
                    else:
                        response += f"**Results:**\n```\n{result_df.to_string()}\n```\n\n"
                else:
                    response += "**No results found.**\n\n"
                
                response += f"**Analysis:**\n{interpretation}"
                
                # Add visualization if requested
                if is_visualization and not result_df.empty:
                    try:
                        print("Visualization requested, attempting to create plot...")
                        # Determine the type of visualization based on the data and query
                        
                        # Check for specific visualization types in the query
                        is_pie_chart = any(word in query.lower() for word in ['pie chart', 'pie graph', 'distribution'])
                        is_histogram = any(word in query.lower() for word in ['histogram', 'distribution of', 'frequency'])
                        is_heatmap = any(word in query.lower() for word in ['heatmap', 'heat map', 'correlation'])
                        is_scatter = any(word in query.lower() for word in ['scatter', 'relationship between', 'correlation'])
                        
                        if len(result_df.columns) >= 2:
                            # Find numeric columns for y-axis
                            numeric_cols = result_df.select_dtypes(include=['number']).columns.tolist()
                            
                            if len(numeric_cols) >= 1 and len(result_df) > 1:
                                # Create appropriate plot based on query and data characteristics
                                if is_pie_chart and len(result_df) <= 20:  # Pie charts work best with limited categories
                                    # For pie charts, we need a category column and a value column
                                    category_col = result_df.columns[0]
                                    value_col = numeric_cols[0] if len(numeric_cols) > 0 else result_df.columns[1]
                                    
                                    fig = px.pie(result_df, names=category_col, values=value_col, 
                                                title="Distribution Analysis",
                                                hole=0.3)  # Use a donut chart for better readability
                                
                                elif is_histogram and len(numeric_cols) > 0:
                                    # For histograms, we need a numeric column
                                    fig = px.histogram(result_df, x=numeric_cols[0], 
                                                    title=f"Distribution of {numeric_cols[0]}",
                                                    nbins=20)
                                
                                elif is_heatmap and len(numeric_cols) >= 2:
                                    # For heatmaps, we need at least 2 numeric columns
                                    # Convert to a correlation matrix if needed
                                    if len(result_df.columns) == len(numeric_cols) and len(numeric_cols) > 2:
                                        # This is likely already a correlation matrix or similar data
                                        fig = px.imshow(result_df, 
                                                    title="Correlation Heatmap",
                                                    color_continuous_scale='RdBu_r',
                                                    aspect="auto")
                                    else:
                                        # Create a correlation matrix from the numeric columns
                                        corr_df = result_df[numeric_cols].corr()
                                        fig = px.imshow(corr_df, 
                                                    title="Correlation Heatmap",
                                                    color_continuous_scale='RdBu_r',
                                                    aspect="auto")
                                
                                elif is_scatter and len(numeric_cols) >= 2:
                                    # For scatter plots, we need at least 2 numeric columns
                                    fig = px.scatter(result_df, x=numeric_cols[0], y=numeric_cols[1],
                                                    title=f"Relationship between {numeric_cols[0]} and {numeric_cols[1]}",
                                                    opacity=0.7)
                                
                                elif 'month' in result_df.columns or 'date' in result_df.columns or 'year' in result_df.columns or any('date' in col.lower() for col in result_df.columns):
                                    # Time series data - use line chart
                                    x_col = result_df.columns[0]
                                    y_cols = numeric_cols[:3]  # Use up to 3 numeric columns
                                    
                                    fig = px.line(result_df, x=x_col, y=y_cols, 
                                                title="Time Series Analysis",
                                                markers=True)
                                else:
                                    # Regular data - use bar chart
                                    x_col = result_df.columns[0]
                                    y_cols = numeric_cols[0]
                                    
                                    fig = px.bar(result_df, x=x_col, y=y_cols, 
                                                title="Data Visualization")
                                
                                # Improve figure layout
                                fig.update_layout(
                                    autosize=True,
                                    width=900,
                                    height=600,
                                    margin=dict(l=50, r=50, b=100, t=100, pad=4),
                                    template="plotly_white",
                                    font=dict(size=14)
                                )
                                
                                # Convert the figure to an image and encode it as base64
                                img_bytes = fig.to_image(format="png", width=900, height=600, scale=2)
                                encoded = base64.b64encode(img_bytes).decode("ascii")
                                img_src = f"data:image/png;base64,{encoded}"
                                
                                # Add the image directly to the response
                                response += f"\n\n<img src='{img_src}' width='100%' />"
                                
                                # Add note about visualization
                                response += "\n\n**A visualization has been generated and is displayed above.**"
                            else:
                                print("Not enough numeric columns or data points for visualization")
                        else:
                            print("Not enough columns for visualization")
                    except Exception as viz_error:
                        print(f"Visualization error: {str(viz_error)}")
                        traceback.print_exc()
            
            except Exception as e:
                response = f"**SQL Query:**\n```sql\n{sql_query}\n```\n\n**Error executing query:** {str(e)}"
            
            conn.close()
            
        except Exception as e:
            response = f"Error processing query: {str(e)}"
    
    else:
        # For non-CSV queries, use the document assistant
        try:
            response = document_assistant.process_query(query)
        except Exception as e:
            response = f"Error processing document query: {str(e)}"
    
    # Calculate processing time
    processing_time = time.time() - start_time
    response += f"\n\n(Query processed in {processing_time:.2f} seconds)"
    
    # Add the response to history
    history.append({"role": "assistant", "content": response})
    
    return "", history

def process_file_upload(files):
    """Process uploaded files and index them"""
    if not files:
        return "No files uploaded"
    
    global current_context
    
    # Clear existing context
    current_context = {
        "file_type": None,
        "file_name": None,
        "table_name": None
    }
    
    file_info = []
    for file in files:
        file_path = file.name
        file_name = os.path.basename(file_path)
        file_ext = os.path.splitext(file_name)[1].lower()
        
        if file_ext == '.csv':
            try:
                # Create table name from filename
                table_name = os.path.splitext(file_name)[0].replace(' ', '_').lower()
                
                # Load CSV into SQLite
                conn = sqlite3.connect(DB_PATH)
                
                # Configure SQLite for faster imports
                conn.execute("PRAGMA synchronous = OFF")
                conn.execute("PRAGMA journal_mode = MEMORY")
                
                # Read the CSV and load it into SQLite
                df = pd.read_csv(file_path)
                df.to_sql('data_tab', conn, if_exists='replace', index=False)
                
                # Update current context
                current_context = {
                    "file_type": "csv",
                    "file_name": file_name,
                    "table_name": "data_tab"  # Always use data_tab as the table name
                }
                
                # Get column info
                cursor = conn.cursor()
                cursor.execute("PRAGMA table_info(data_tab);")
                columns = [f"{col[1]} ({col[2]})" for col in cursor.fetchall()]
                
                # Get row count
                cursor.execute("SELECT COUNT(*) FROM data_tab;")
                row_count = cursor.fetchone()[0]
                
                conn.close()
                
                file_info.append("✅ CSV File Successfully Loaded")
                file_info.append(f"📊 Table Name: data_tab")
                file_info.append(f"📄 Source File: {file_name}")
                file_info.append(f"📈 Total Rows: {row_count:,}")
                file_info.append(f"📋 Columns: {', '.join(columns)}")
                
            except Exception as e:
                file_info.append(f"❌ Error loading CSV {file_name}: {str(e)}")
        
        else:
            # Process PDF or other document types
            try:
                result = document_assistant.upload_document(file_path)
                
                # Update current context
                current_context = {
                    "file_type": "pdf",
                    "file_name": file_name,
                    "table_name": None
                }
                
                file_info.append("✅ Document Successfully Processed")
                file_info.append(f"📄 File: {file_name}")
                file_info.append(f"📚 Chunks: {result['chunks']}")
                file_info.append(result['message'])
            except Exception as e:
                file_info.append(f"❌ Error processing document {file_name}: {str(e)}")
    
    return "\n".join(file_info)

def list_documents():
    """List all indexed documents"""
    info_list = []
    
    # Check for CSV data
    try:
        conn = sqlite3.connect(DB_PATH)
        cursor = conn.cursor()
        cursor.execute("SELECT name FROM sqlite_master WHERE type='table';")
        tables = cursor.fetchall()
        
        if tables:
            info_list.append("📊 CSV Data Tables:")
            for table in tables:
                # Get column info
                cursor.execute(f"PRAGMA table_info({table[0]});")
                columns = [col[1] for col in cursor.fetchall()]
                
                # Get row count
                cursor.execute(f"SELECT COUNT(*) FROM {table[0]};")
                row_count = cursor.fetchone()[0]
                
                info_list.append(f"- {table[0]} ({row_count:,} rows, {len(columns)} columns)")
        
        conn.close()
    except Exception as e:
        info_list.append(f"Error accessing CSV data: {str(e)}")
    
    # Check for indexed documents
    docs = document_assistant.get_all_documents()
    if docs:
        info_list.append("\n📑 Indexed Documents:")
        for doc in docs:
            info_list.append(f"- {doc['filename']} (ID: {doc['id']})")
    
    if not info_list:
        return "No data or documents loaded yet"
    
    return "\n".join(info_list)

def clear_context():
    """Clear the current context and chat history"""
    global current_context
    current_context = {
        "file_type": None,
        "file_name": None,
        "table_name": None
    }
    return None

def process_voice_input(audio_path):
    """Process voice input and return transcribed text"""
    if audio_path is None:
        return "No audio recorded"
    
    try:
        # Initialize recognizer
        r = sr.Recognizer()
        
        # Load the audio file
        with sr.AudioFile(audio_path) as source:
            # Read the audio data
            audio_data = r.record(source)
            
            # Recognize speech using Google Speech Recognition
            text = r.recognize_google(audio_data)
            
            return text
    except sr.UnknownValueError:
        return "Could not understand audio"
    except sr.RequestError as e:
        return f"Error with speech recognition service: {e}"
    except Exception as e:
        return f"Error processing audio: {str(e)}"

def text_to_speech_output(text):
    """Convert text to speech"""
    if not text or len(text) == 0:
        return None
    
    # Extract the last assistant message
    last_message = None
    for msg in reversed(text):
        if msg["role"] == "assistant":
            last_message = msg["content"]
            break
    
    if not last_message:
        return None
    
    try:
        # Clean the text (remove markdown and HTML)
        clean_text = re.sub(r'<.*?>', '', last_message)  # Remove HTML tags
        clean_text = re.sub(r'\*\*(.*?)\*\*', r'\1', clean_text)  # Remove bold markdown
        clean_text = re.sub(r'\n\n', ' ', clean_text)  # Replace double newlines with space
        clean_text = re.sub(r'```.*?```', 'Code block removed for speech.', clean_text, flags=re.DOTALL)  # Replace code blocks
        
        # Create a temporary file
        temp_file = tempfile.NamedTemporaryFile(delete=False, suffix=".mp3")
        temp_file.close()
        
        # Generate speech
        tts = gTTS(text=clean_text, lang='en', slow=False)
        tts.save(temp_file.name)
        
        return temp_file.name
    except Exception as e:
        print(f"Error generating speech: {str(e)}")
        return None

def create_test_visualization():
    """Create a test visualization to verify plotting works"""
    # Create sample data
    data = pd.DataFrame({
        'Month': ['Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun'],
        'Value': [10, 15, 13, 17, 20, 25]
    })
    
    # Create a simple bar chart
    fig = px.bar(data, x='Month', y='Value', title='Test Visualization')
    
    # Configure the figure
    fig.update_layout(
        autosize=True,
        width=800,
        height=500
    )
    
    return fig

def create_test_html_visualization():
    """Create a test HTML visualization to verify plotting works"""
    # Create sample data
    data = pd.DataFrame({
        'Month': ['Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun'],
        'Value': [10, 15, 13, 17, 20, 25]
    })
    
    # Create a simple bar chart
    fig = px.bar(data, x='Month', y='Value', title='Test Visualization')
    
    # Convert to HTML with CDN-hosted plotly.js
    html = pio.to_html(fig, full_html=False, include_plotlyjs='cdn')
    
    return html

def flush_databases():
    """Flush ChromaDB and SQLite databases"""
    result = []
    
    # Flush SQLite database
    try:
        conn = sqlite3.connect(DB_PATH)
        cursor = conn.cursor()
        
        # Get all tables
        cursor.execute("SELECT name FROM sqlite_master WHERE type='table';")
        tables = cursor.fetchall()
        
        # Drop all tables
        for table in tables:
            cursor.execute(f"DROP TABLE IF EXISTS {table[0]};")
        
        conn.commit()
        conn.close()
        
        result.append("✅ SQLite database cleared successfully")
    except Exception as e:
        result.append(f"❌ Error clearing SQLite database: {str(e)}")
    
    # Flush ChromaDB by resetting the document assistant
    try:
        document_assistant.reset_database()
        result.append("✅ ChromaDB cleared successfully")
    except Exception as e:
        result.append(f"❌ Error clearing ChromaDB: {str(e)}")
    
    # Reset current context
    global current_context
    current_context = {
        "file_type": None,
        "file_name": None,
        "table_name": None
    }
    
    return "\n".join(result)

# Create Gradio interface
with gr.Blocks(title="AI Document Analysis & Voice Assistant") as demo:
    gr.Markdown("# 🤖 AI Document Analysis & Voice Assistant")
    gr.Markdown("Upload documents, ask questions, and get voice responses!")
    
    with gr.Tab("Chat"):
        # Use a custom CSS to ensure images are displayed properly
        gr.HTML("""
        <style>
        .chatbot-container img {
            max-width: 100%;
            height: auto;
            display: block;
            margin: 10px 0;
        }
        </style>
        """)
        
        chatbot = gr.Chatbot(height=500, type="messages", elem_classes="chatbot-container")
        
        with gr.Row():
            with gr.Column(scale=8):
                msg = gr.Textbox(
                    placeholder="Ask a question about your documents...",
                    show_label=False
                )
            with gr.Column(scale=1):
                voice_btn = gr.Button("🎤")
        
        with gr.Row():
            submit_btn = gr.Button("Submit")
            clear_btn = gr.Button("Clear")
            clear_context_btn = gr.Button("Clear Context")
        
        audio_output = gr.Audio(label="Voice Response", type="filepath")
        
        # Voice input
        voice_input = gr.Audio(
            label="Voice Input", 
            type="filepath",
            visible=False
        )
        
        # Event handlers
        submit_btn.click(
            process_text_query, 
            inputs=[msg, chatbot], 
            outputs=[msg, chatbot]
        )
        
        msg.submit(
            process_text_query, 
            inputs=[msg, chatbot], 
            outputs=[msg, chatbot]
        )
        
        clear_btn.click(lambda: None, None, [chatbot], queue=False)
        clear_context_btn.click(clear_context, inputs=[], outputs=[chatbot])
        
        voice_btn.click(
            lambda: gr.update(visible=True),
            None,
            voice_input
        )
        
        voice_input.change(
            process_voice_input,
            inputs=[voice_input],
            outputs=[msg]
        )
        
        # Add TTS functionality
        tts_btn = gr.Button("🔊 Speak Response")
        tts_btn.click(
            text_to_speech_output,
            inputs=[chatbot],
            outputs=[audio_output]
        )
    
    with gr.Tab("Document Upload"):
        with gr.Row():
            file_upload = gr.File(
                label="Upload Documents",
                file_types=[".pdf", ".txt", ".docx", ".csv", ".xlsx"],
                file_count="multiple"
            )
            flush_db_btn_doc = gr.Button("🗑️ Flush All Databases", variant="stop")
        
        upload_button = gr.Button("Process & Index Documents")
        upload_output = gr.Textbox(label="Upload Status")
        
        upload_button.click(
            process_file_upload,
            inputs=[file_upload],
            outputs=[upload_output]
        )
        
        flush_db_btn_doc.click(
            flush_databases,
            inputs=[],
            outputs=[upload_output]
        )
        
        list_docs_button = gr.Button("List Indexed Documents")
        docs_output = gr.Textbox(label="Indexed Documents")
        
        list_docs_button.click(
            list_documents,
            inputs=[],
            outputs=[docs_output]
        )
    
    with gr.Tab("Settings"):
        with gr.Row():
            gr.Markdown("## Database Management")
            flush_db_btn = gr.Button("🗑️ Flush All Databases", variant="stop", scale=1)
        
        flush_result = gr.Textbox(label="Flush Result")
        
        flush_db_btn.click(
            flush_databases,
            inputs=[],
            outputs=[flush_result]
        )
        
        gr.Markdown("## System Settings")
        api_key = gr.Textbox(
            label="Groq API Key",
            placeholder="Enter your Groq API key",
            type="password",
            value=os.getenv("GROQ_API_KEY", "")
        )
        save_btn = gr.Button("Save Settings")
        
        def save_settings(key):
            os.environ["GROQ_API_KEY"] = key
            return "Settings saved!"
        
        save_btn.click(
            save_settings,
            inputs=[api_key],
            outputs=[gr.Textbox(label="Status")]
        )
        
        gr.Markdown("## Debugging")
        test_viz_btn = gr.Button("Test Visualization")
        test_viz_output = gr.HTML(label="Test Visualization")
        
        test_viz_btn.click(
            create_test_html_visualization,
            inputs=[],
            outputs=[test_viz_output]
        )

# Launch the app
if __name__ == "__main__":
    demo.launch()