src/app.py

"""
FILE 1: src/app.py — Flask Server (Entry Point)
=================================================
IMPORTS FROM: model.py (init_model, predict_qa), scraper.py (scrape_url)
SERVES:       templates/index.html via route "/"
CALLED BY:    static/js/main.js via fetch() to /api/scrape and /api/predict

Routes:
  GET  /             → serves index.html (the UI)
  POST /api/scrape   → receives {url}, calls scraper.py, returns scraped text
  POST /api/predict  → receives {question, context}, calls model.py, returns answer
"""

import os
import logging
from flask import Flask, render_template, request, jsonify

from model import init_model, predict_qa
from scraper import scrape_url

FORMAT = "%(asctime)-15s [%(levelname)s] %(message)s"
logging.basicConfig(format=FORMAT, level=logging.INFO)
logger = logging.getLogger(__name__)

app = Flask(
    __name__,
    template_folder=os.path.join(os.path.dirname(__file__), "..", "templates"),
    static_folder=os.path.join(os.path.dirname(__file__), "..", "static"),
)


@app.route("/")
def index():
    """Serve the main UI page."""
    return render_template("index.html")


@app.route("/api/scrape", methods=["POST"])
def api_scrape():
    """
    Called by: main.js → scrapeURL()
    Input:    { "url": "https://amazon.in/..." }
    Output:   { "title", "context", "source", "char_count" }
    Calls:    scraper.py → scrape_url()
    """
    data = request.get_json()
    url = data.get("url", "").strip()
    if not url:
        return jsonify({"error": "URL is required."}), 400
    try:
        result = scrape_url(url)
        if result.get("error"):
            return jsonify(result), 400
        return jsonify(result)
    except Exception as e:
        logger.exception("Scraping failed")
        return jsonify({"error": str(e)}), 500


@app.route("/api/predict", methods=["POST"])
def api_predict():
    """
    Called by: main.js → doAsk()
    Input:    { "question": "What is the battery?", "context": "Samsung Galaxy..." }
    Output:   { "answer", "confidence", "confidence_pct", "confidence_level",
                "tokens", "answer_start_char", "answer_end_char", ... }
    Calls:    model.py → predict_qa()
    """
    data = request.get_json()
    question = data.get("question", "").strip()
    context = data.get("context", "").strip()
    if not question or not context:
        return jsonify({"error": "Both question and context are required."}), 400
    if len(context) < 20:
        return jsonify({"error": "Context too short. Need at least a few sentences."}), 400
    try:
        result = predict_qa(question, context)
        return jsonify(result)
    except Exception as e:
        logger.exception("Prediction failed")
        return jsonify({"error": str(e)}), 500


# Initialize the BERT model at import time so gunicorn workers have it ready.
# This runs both when gunicorn imports the app AND when you run `python app.py` directly.
logger.info("Initializing BERT QA model...")
init_model()
logger.info("Model ready.")

if __name__ == "__main__":
    logger.info("Server starting on http://localhost:5000")
    app.run(host="0.0.0.0", port=int(os.environ.get("PORT", 5000)), debug=False)