import logging import io import base64 from flask import Flask, render_template, request, jsonify from PIL import Image from src.config import MAX_CONTENT_LENGTH, MODEL_PATH from src.inference import get_or_load_resources, predict_review, generate_explanation # Set up logging configuration logging.basicConfig( level=logging.INFO, format='%(asctime)s [%(levelname)s] %(name)s: %(message)s' ) logger = logging.getLogger(__name__) app = Flask(__name__) app.config['MAX_CONTENT_LENGTH'] = MAX_CONTENT_LENGTH # Global status tracker models_loaded = False # Warm up model weights, download sentiment lexicons, and cache-boot clients immediately at module import try: logger.info("Initializing SentinelReview AI backend resources at startup...") get_or_load_resources() models_loaded = True logger.info("SentinelReview AI initialization complete. System fully operational.") except Exception as e: logger.critical(f"Failed to initialize server models during startup: {e}", exc_info=True) models_loaded = False @app.route('/') def index(): """Serves the main dashboard application interface.""" return render_template('index.html', models_status=models_loaded) @app.route('/analyze', methods=['POST']) def analyze(): """ API Endpoint: Audits a review text and context image. Performs classification and requests text explanations. """ if not models_loaded: return jsonify({"error": "Platform models failed to load. Please audit server logs."}), 503 review_text = request.form.get('review_text', '').strip() if not review_text: return jsonify({"error": "No text review content was provided."}), 400 # 1. Process context image (if uploaded) image_b64 = None if 'review_image' in request.files and request.files['review_image'].filename != '': image_file = request.files['review_image'] try: image = Image.open(image_file) image.thumbnail((400, 400)) buffered = io.BytesIO() image.save(buffered, format="JPEG") image_b64 = base64.b64encode(buffered.getvalue()).decode('utf-8') logger.info("Context image successfully parsed and encoded.") except Exception as e: logger.warning(f"Failed to process context image: {e}") # 2. Run core classification prediction try: logger.info(f"Auditing review text (Length: {len(review_text)} chars)...") results = predict_review(review_text) # 3. Retrieve explanation explanation = generate_explanation(review_text, results['classification']) # 4. Compile audit response return jsonify({ "classification": results['classification'], "confidence": f"{int(results['confidence'] * 100)}%", "explanation": explanation, "image": image_b64, "metrics": results['metrics'] }) except ValueError as val_err: logger.error(f"Validation error: {val_err}") return jsonify({"error": str(val_err)}), 400 except Exception as err: logger.error(f"Inference crash during review classification: {err}", exc_info=True) return jsonify({"error": f"Internal classification pipeline error: {str(err)}"}), 500 if __name__ == '__main__': app.run(host='0.0.0.0', port=7860)