app.py

# app.py - Flask Backend
from flask import Flask, request, jsonify, send_from_directory
import google.generativeai as genai
from dotenv import load_dotenv
import os
from flask_cors import CORS
import markdown2
import re
from gtts import gTTS
import uuid

# Load environment variables
load_dotenv()

# Define paths and create static audio directory if it doesn't exist
AUDIO_FOLDER = os.path.join('static', 'audio')
if not os.path.exists(AUDIO_FOLDER):
    os.makedirs(AUDIO_FOLDER)

# Initialize Flask app
app = Flask(__name__, static_folder='static')
CORS(app)  # Enable CORS for all routes

# Configure Gemini with a system instruction
system_instruction_text = """
You are a helpful, friendly, and informative AI assistant named Athspi.
Your goal is to provide clear, concise, and natural-sounding answers to user queries.
When you respond:
- Use clear and simple language.
- Avoid overly complex sentence structures that might be hard to read aloud.
- Keep the user engaged and offer follow-up questions or related topics where appropriate.
- Ensure your responses are suitable for text-to-speech conversion.
- Provide factual and accurate information.
- If the user asks for audio or to speak the response, include 🔊 at the start of your response.
- For coding questions, provide well-formatted code blocks with syntax highlighting.
"""

genai.configure(api_key=os.getenv("GEMINI_API_KEY"))
model = genai.GenerativeModel(
    'gemini-2.5-flash',
    system_instruction=system_instruction_text
)

def convert_markdown_to_html(text):
    html = markdown2.markdown(text, extras=["fenced-code-blocks", "tables"])
    html = re.sub(r'<pre><code(.*?)>', r'<pre class="code-block"><code\1>', html)
    html = re.sub(r'\*\*(.*?)\*\*', r'<strong>\1</strong>', html)
    html = re.sub(r'\*(.*?)\*', r'<em>\1</em>', html)
    return html

def detect_audio_request(text):
    """Detect if user is requesting audio"""
    audio_keywords = ['audio', 'speak', 'say it', 'read aloud', 'hear', 'listen']
    return any(keyword in text.lower() for keyword in audio_keywords)

def generate_audio_file(text):
    """Generate audio file from text and return filename"""
    cleaned_text = re.sub(r'[\*_`#]', '', text)
    cleaned_text = re.sub(r'\s+', ' ', cleaned_text).strip()

    if not cleaned_text:
        return None

    filename = f"{uuid.uuid4()}.mp3"
    filepath = os.path.join(AUDIO_FOLDER, filename)

    tts = gTTS(text=cleaned_text, lang='en', slow=False)
    tts.save(filepath)

    return filename

@app.route('/chat', methods=['POST'])
def chat():
    try:
        data = request.json
        user_message = data.get('message')
        
        if not user_message:
            return jsonify({"error": "No message provided"}), 400
        
        # Detect if user is requesting audio
        audio_requested = detect_audio_request(user_message)
        
        response = model.generate_content(user_message)
        plain_text_response = response.text
        
        # Generate audio if requested
        audio_url = None
        if audio_requested:
            # Add audio indicator if not already present
            if not plain_text_response.startswith("🔊"):
                plain_text_response = "🔊 " + plain_text_response
            
            # Generate audio file
            audio_filename = generate_audio_file(plain_text_response)
            if audio_filename:
                audio_url = f"/static/audio/{audio_filename}"
        
        html_response = convert_markdown_to_html(plain_text_response)
        
        return jsonify({
            "response_html": html_response,
            "response_text": plain_text_response,
            "audio_url": audio_url
        })
        
    except Exception as e:
        app.logger.error(f"Chat Error: {e}")
        return jsonify({"error": str(e)}), 500

@app.route('/generate-audio', methods=['POST'])
def generate_audio():
    try:
        data = request.json
        text_to_speak = data.get('text')

        if not text_to_speak:
            return jsonify({"error": "No text provided"}), 400

        cleaned_text = re.sub(r'[\*_`#]', '', text_to_speak)
        cleaned_text = re.sub(r'\s+', ' ', cleaned_text).strip()

        if not cleaned_text:
            return jsonify({"error": "Text became empty after cleaning"}), 400

        filename = f"{uuid.uuid4()}.mp3"
        filepath = os.path.join(AUDIO_FOLDER, filename)

        tts = gTTS(text=cleaned_text, lang='en', slow=False)
        tts.save(filepath)

        audio_url = f"/static/audio/{filename}"
        return jsonify({"audio_url": audio_url})

    except Exception as e:
        app.logger.error(f"Audio Generation Error: {e}")
        return jsonify({"error": str(e)}), 500

@app.route('/static/audio/<filename>')
def serve_audio(filename):
    return send_from_directory(AUDIO_FOLDER, filename)

@app.route('/')
def serve_index():
    return send_from_directory('static', 'index.html')

@app.route('/<path:path>')
def serve_static(path):
    return send_from_directory('static', path)

if __name__ == '__main__':
    app.run(host="0.0.0.0", port=7860)