Upload 3 files

backend/app.py

@@ -0,0 +1,151 @@
+import os
+from typing import List
+import numpy as np
+import pandas as pd
+from PIL import Image
+import tensorflow as tf
+from tensorflow.keras.models import load_model
+from tensorflow.keras.preprocessing import image
+from tensorflow.keras.preprocessing.image import load_img
+from tensorflow.keras.preprocessing.image import img_to_array
+from models.skin_tone.skin_tone_knn import identify_skin_tone
+from flask import Flask, request
+from flask_restful import Api, Resource, reqparse, abort
+import werkzeug
+from models.recommender.rec import recs_essentials, makeup_recommendation
+import base64
+from io import BytesIO
+import logging
+
+app = Flask(__name__)
+api = Api(app)
+
+# Set up logging
+logging.basicConfig(level=logging.DEBUG)
+logger = logging.getLogger(__name__)
+
+class_names1 = ['Dry_skin', 'Normal_skin', 'Oil_skin']
+class_names2 = ['Low', 'Moderate', 'Severe']
+skin_tone_dataset = 'models/skin_tone/skin_tone_dataset.csv'
+
+def get_model():
+    global model1, model2
+    model1 = load_model('./models/skin_model')
+    print('Model 1 loaded')
+    model2 = load_model('./models/acne_model')
+    print("Model 2 loaded!")
+
+def load_image(img_path):
+    img = image.load_img(img_path, target_size=(224, 224))
+    img_tensor = image.img_to_array(img)
+    img_tensor = np.expand_dims(img_tensor, axis=0)
+    img_tensor /= 255.
+    return img_tensor
+
+def prediction_skin(img_path):
+    new_image = load_image(img_path)
+    pred1 = model1.predict(new_image)
+    if len(pred1[0]) > 1:
+        pred_class1 = class_names1[tf.argmax(pred1[0])]
+    else:
+        pred_class1 = class_names1[int(tf.round(pred1[0]))]
+    return pred_class1
+
+def prediction_acne(img_path):
+    new_image = load_image(img_path)
+    pred2 = model2.predict(new_image)
+    if len(pred2[0]) > 1:
+        pred_class2 = class_names2[tf.argmax(pred2[0])]
+    else:
+        pred_class2 = class_names2[int(tf.round(pred2[0]))]
+    return pred_class2
+
+get_model()
+
+# Parsing arguments for image and recommendations
+img_put_args = reqparse.RequestParser()
+img_put_args.add_argument("file", help="Please provide a valid image file", required=True)
+
+rec_args = reqparse.RequestParser()
+rec_args.add_argument("tone", type=int, help="Argument required", required=True)
+rec_args.add_argument("type", type=str, help="Argument required", required=True)
+rec_args.add_argument("features", type=dict, help="Argument required", required=True)
+
+# Recommendation Class
+class Recommendation(Resource):
+    logger.info(f"Received recommendation request before --------")
+    def put(self):
+        args = rec_args.parse_args()
+        
+        # Log the incoming recommendation request
+        logger.info(f"Received recommendation request with data: {args}")
+        
+        features = args['features']
+        tone = args['tone']
+        skin_type = args['type'].lower()
+        skin_tone = 'light to medium'
+        
+        # Adjust skin tone based on the tone input
+        if tone <= 2:
+            skin_tone = 'fair to light'
+        elif tone >= 4:
+            skin_tone = 'medium to dark'
+        
+        # Log the skin tone and type
+        logger.info(f"Skin tone: {skin_tone}, Skin type: {skin_type}")
+        
+        fv = []
+        for key, value in features.items():
+            fv.append(int(value))
+
+        # Log the features sent for recommendation
+        logger.info(f"Features: {fv}")
+        
+        try:
+            general = recs_essentials(fv, None)
+            makeup = makeup_recommendation(skin_tone, skin_type)
+            
+            # Log the recommendation data being returned
+            logger.info(f"Generated recommendations: General: {general}, Makeup: {makeup}")
+            
+            return {'general': general, 'makeup': makeup}
+        
+        except Exception as e:
+            logger.error(f"Error during recommendation generation: {str(e)}")
+            return {'error': 'Error processing recommendations'}, 400
+
+
+
+# Skin Metrics Class
+class SkinMetrics(Resource):
+    def put(self):
+        args = img_put_args.parse_args()
+        
+        # Log the incoming image request
+        logger.info(f"Received image for skin metrics analysis: {args}")
+        
+        file = args['file']
+        starter = file.find(',')
+        image_data = file[starter+1:]
+        
+        image_data = bytes(image_data, encoding="ascii")
+        im = Image.open(BytesIO(base64.b64decode(image_data+b'==')))
+
+        filename = 'image.png'
+        file_path = os.path.join('./static', filename)
+        im.save(file_path)
+
+        skin_type = prediction_skin(file_path).split('_')[0]
+        acne_type = prediction_acne(file_path)
+        tone = identify_skin_tone(file_path, dataset=skin_tone_dataset)
+        
+        # Log the predictions for skin type, acne type, and skin tone
+        logger.info(f"Predicted skin type: {skin_type}, acne type: {acne_type}, tone: {tone}")
+        
+        return {'type': skin_type, 'tone': str(tone), 'acne': acne_type}, 200
+
+api.add_resource(SkinMetrics, "/upload")
+api.add_resource(Recommendation, "/recommend")
+
+if __name__ == "__main__":
+    app.run(debug=False)

backend/requirements.txt

@@ -1,13 +1,13 @@
-numpy
-pandas
-matplotlib
-scikit-learn
+numpy==1.24.3
+pandas==1.5.0
+matplotlib==3.9.2
+scikit-learn==1.3.0
 Flask==2.0.2
 Flask-RESTful==0.3.9
-tensorflow
+tensorflow-macos==2.13.0
 opencv-python
-Werkzeug
+Werkzeug==2.1.2
 Jinja2
 click
 reparse==3.0
-Pillow
+Pillow==9.3.0