import os os.environ['MPLCONFIGDIR'] = '/tmp/matplotlib' from flask import Flask, render_template, request, redirect, url_for, session, send_file, jsonify from flask_sqlalchemy import SQLAlchemy from flask_migrate import Migrate import threading import tensorflow as tf import numpy as np from PIL import Image import pickle import io import matplotlib.pyplot as plt from reportlab.lib.pagesizes import A4 from reportlab.lib import colors from reportlab.pdfgen import canvas from reportlab.lib.units import inch from datetime import datetime import logging from flask_mail import Mail, Message import base64 from sendgrid import SendGridAPIClient from sendgrid.helpers.mail import (Mail, Attachment, FileContent, FileName, FileType, Disposition) app = Flask(__name__) app.secret_key = "e3f6f40bb8b2471b9f07c4025d845be9" # Database configuration app.config['SQLALCHEMY_DATABASE_URI'] = 'sqlite:////tmp/snapsin.db' app.config['SQLALCHEMY_TRACK_MODIFICATIONS'] = False db = SQLAlchemy(app) migrate = Migrate(app, db) # Mail configuration app.config['MAIL_SERVER'] = 'smtp.gmail.com' app.config['MAIL_PORT'] = 465 app.config['MAIL_USERNAME'] = os.environ.get('MAIL_USERNAME') app.config['MAIL_PASSWORD'] = os.environ.get('MAIL_PASSWORD') app.config['MAIL_USE_TLS'] = False app.config['MAIL_USE_SSL'] = True mail = Mail(app) MODEL_PATH = "skin_lesion_model.h5" HISTORY_PATH = "training_history.pkl" PLOT_PATH = "/tmp/static/training_plot.png" LOGO_PATH = "static/logo.jpg" FORM_TEMPLATE = "form.html" # Use form.html for both input and results IMG_SIZE = (224, 224) CONFIDENCE_THRESHOLD = 0.30 label_map = { 0: "Melanoma", 1: "Melanocytic nevus", 2: "Basal cell carcinoma", 3: "Actinic keratosis", 4: "Benign keratosis", 5: "Dermatofibroma", 6: "Vascular lesion", 7: "Squamous cell carcinoma" } recommendations = { "Melanoma": { "solutions": ["Consult a dermatologist immediately.", "Surgical removal is typically required.", "Regular follow-up and screening for metastasis."], "medications": ["Interferon alfa-2b", "Vemurafenib", "Dacarbazine"] }, "Melanocytic nevus": { "solutions": ["Usually benign and requires no treatment.", "Monitor for any change in shape or color."], "medications": ["No medication necessary unless changes occur."] }, "Basal cell carcinoma": { "solutions": ["Surgical excision or Mohs surgery.", "Topical treatments if superficial.", "Radiation in select cases."], "medications": ["Imiquimod cream", "Fluorouracil cream", "Vismodegib"] }, "Actinic keratosis": { "solutions": ["Cryotherapy or topical treatments.", "Avoid prolonged sun exposure.", "Use of sunscreen regularly."], "medications": ["Fluorouracil", "Imiquimod", "Diclofenac gel"] }, "Benign keratosis": { "solutions": ["Generally harmless and often left untreated.", "Can be removed for cosmetic reasons."], "medications": ["No medication required unless infected."] }, "Dermatofibroma": { "solutions": ["Benign skin growth, no treatment needed.", "Surgical removal if painful or for cosmetic reasons."], "medications": ["No medication needed."] }, "Vascular lesion": { "solutions": ["Treatment depends on type (e.g., hemangioma).", "Laser therapy is commonly used.", "Observation if no complications."], "medications": ["Beta-blockers (e.g., propranolol for hemangioma)"] }, "Squamous cell carcinoma": { "solutions": ["Surgical removal is standard.", "Follow-up for recurrence or metastasis.", "Avoid sun exposure and use sunscreen."], "medications": ["Fluorouracil", "Cisplatin", "Imiquimod"] }, "Low confidence": { "solutions": ["The image is not confidently classified.", "Please upload a clearer image or consult a doctor."], "medications": ["Not available due to low confidence."] }, "Unknown": {"solutions": ["No specific guidance available."], "medications": ["N/A"]} } logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s') logger = logging.getLogger(__name__) class User(db.Model): id = db.Column(db.Integer, primary_key=True) name = db.Column(db.String(100), nullable=False) email = db.Column(db.String(120), unique=True, nullable=False) scans = db.relationship('Scan', backref='user', lazy=True) class Scan(db.Model): id = db.Column(db.Integer, primary_key=True) user_id = db.Column(db.Integer, db.ForeignKey('user.id'), nullable=False) patient_name = db.Column(db.String(100), nullable=False) patient_gender = db.Column(db.String(20), nullable=False) patient_age = db.Column(db.Integer, nullable=False) prediction = db.Column(db.String(100), nullable=False) confidence = db.Column(db.String(20), nullable=False) timestamp = db.Column(db.DateTime, default=datetime.utcnow) image_filename = db.Column(db.String(100), nullable=False) model = None model_load_error = None def load_model(): global model, model_load_error try: if os.path.exists(MODEL_PATH): model = tf.keras.models.load_model(MODEL_PATH, compile=False) logger.info("Model loaded successfully") else: model_load_error = f"Model file {MODEL_PATH} not found" logger.error(model_load_error) except Exception as e: model_load_error = f"Model deserialization error: {e}" logger.error(f"Failed to load model: {e}") load_model() if os.path.exists(HISTORY_PATH): try: with open(HISTORY_PATH, "rb") as f: history_dict = pickle.load(f) if "accuracy" in history_dict and "val_accuracy" in history_dict: os.makedirs("/tmp/static", exist_ok=True) plt.figure() plt.plot(history_dict['accuracy'], label='Train Accuracy') plt.plot(history_dict['val_accuracy'], label='Val Accuracy') plt.xlabel('Epochs') plt.ylabel('Accuracy') plt.title('Training History') plt.legend() plt.grid(True) plt.savefig(PLOT_PATH) plt.close() except Exception as e: logger.warning(f"Training history load error: {e}") def preprocess_image(image_bytes): image = Image.open(io.BytesIO(image_bytes)).convert("RGB") image = image.resize(IMG_SIZE) image_array = tf.keras.utils.img_to_array(image) return np.expand_dims(image_array, axis=0) / 255.0 # REPLACE your old function with this new one def send_email_async(app_context, report_data): with app_context: try: # 1. Create the PDF report pdf_path = f"/tmp/report_{report_data['scan_id']}.pdf" generate_pdf(report_data, pdf_path) # 2. Create the email message message = Mail( from_email='snapskinofficial@gmail.com', # IMPORTANT: Use the email you verified on SendGrid to_emails=report_data['email'], subject='Your SnapSkin Diagnostic Report', html_content=f"""
""" ) # 3. Read the PDF and attach it to the email with open(pdf_path, 'rb') as f: data = f.read() encoded_file = base64.b64encode(data).decode() attachedFile = Attachment( FileContent(encoded_file), FileName(f"SnapSkin_Report_{report_data['scan_id']}.pdf"), FileType('application/pdf'), Disposition('attachment') ) message.attachment = attachedFile # 4. Use the API key from secrets to send the email sg = SendGridAPIClient(os.environ.get('SENDGRID_API_KEY')) response = sg.send(message) # 5. Clean up the created PDF file os.remove(pdf_path) logger.info(f"Report sent via SendGrid, status code: {response.status_code}") except Exception as e: logger.error(f"Failed to send email via SendGrid: {e}") def generate_pdf(report, filepath): c = canvas.Canvas(filepath, pagesize=A4) width, height = A4 y = height - 60 c.setFillColor(colors.Color(0.98, 0.98, 0.99, alpha=1)) c.rect(0, 0, width, height, fill=1, stroke=0) c.setFillColor(colors.Color(0.94, 0.96, 0.98, alpha=1)) c.rect(0, height-120, width, 120, fill=1, stroke=0) if os.path.exists(LOGO_PATH): c.drawImage(LOGO_PATH, 67, y-23, width=46, height=46, preserveAspectRatio=True, mask='auto') c.setFont("Helvetica-Bold", 22) c.setFillColor(colors.Color(0.2, 0.2, 0.2, alpha=1)) c.drawCentredString(width / 2, y + 5, "SnapSkin Diagnosis Report") c.setFont("Helvetica", 11) c.setFillColor(colors.Color(0.5, 0.5, 0.5, alpha=1)) c.drawCentredString(width / 2, y - 15, "Dermatological Analysis") c.setStrokeColor(colors.Color(0.8, 0.8, 0.8, alpha=1)) c.line(80, y - 35, width - 80, y - 35) y -= 80 def professional_section_box(title, fields, extra_gap=20): nonlocal y box_height = len(fields) * 20 + 40 c.setFillColor(colors.white) c.roundRect(40, y - box_height, width - 80, box_height, 10, fill=1, stroke=1) c.setStrokeColor(colors.Color(0.9, 0.9, 0.9, alpha=1)) c.setFillColor(colors.Color(0.95, 0.95, 0.95, alpha=1)) c.roundRect(40, y - 30, width - 80, 30, 10, fill=1, stroke=0) c.setFont("Helvetica-Bold", 12) c.setFillColor(colors.Color(0.3, 0.3, 0.3, alpha=1)) c.drawString(55, y - 20, title) y -= 45 for label, val in fields.items(): c.setFont("Helvetica-Bold", 9) c.setFillColor(colors.Color(0.4, 0.4, 0.4, alpha=1)) c.drawString(55, y, f"{label}:") c.setFont("Helvetica", 9) c.setFillColor(colors.Color(0.2, 0.2, 0.2, alpha=1)) c.drawString(150, y, str(val)) y -= 20 y -= extra_gap professional_section_box("Patient Information", { "Name": report["name"], "Email": report["email"], "Gender": report["gender"], "Age": f"{report['age']} years" }) confidence_val = float(report["confidence"].replace('%', '')) confidence_text = f"{report['confidence']} ({'High' if confidence_val > 85 else 'Moderate' if confidence_val > 70 else 'Low'} Confidence)" professional_section_box("Diagnostic Results", { "Condition": report["prediction"], "Confidence": confidence_text, "Notes": report.get("message", "No additional notes") }) treatment = recommendations.get(report["prediction"], recommendations["Unknown"]) professional_section_box("Treatment Recommendations", {f"{i+1}. {line}": "" for i, line in enumerate(treatment["solutions"])}) professional_section_box("Medication Guidelines", {f"{i+1}. {line}": "" for i, line in enumerate(treatment["medications"])}) c.setFillColor(colors.Color(0.98, 0.98, 0.98, alpha=1)) c.roundRect(40, 40, width - 80, 70, 10, fill=1, stroke=1) c.setStrokeColor(colors.Color(0.9, 0.9, 0.9, alpha=1)) c.setFont("Helvetica-Bold", 10) c.setFillColor(colors.Color(0.4, 0.4, 0.4, alpha=1)) c.drawString(50, 95, "Medical Disclaimer") c.setFont("Helvetica", 8) disclaimer = "This report is AI-generated for preliminary assessment. It is not a substitute for professional medical advice. Please consult a qualified healthcare provider." c.drawString(50, 80, disclaimer[:110]) c.drawString(50, 70, disclaimer[110:]) c.save() @app.route("/") def home(): return redirect(url_for("form")) @app.route("/form") def form(): if model_load_error: return render_template(FORM_TEMPLATE, history_plot="/training_plot.png", result={ "prediction": "Error", "confidence": "N/A", "message": f"Model loading failed: {model_load_error}", "email_status": "N/A" }) return render_template(FORM_TEMPLATE, history_plot="/training_plot.png") @app.route("/training_plot.png") def training_plot(): return send_file(PLOT_PATH, mimetype="image/png") if os.path.exists(PLOT_PATH) else ("", 404) @app.route("/uploads/