Update main.py

main.py

@@ -1,7 +1,7 @@
 import os
 os.environ["TRANSFORMERS_CACHE"] = "/tmp/huggingface_cache"  # Set cache directory to a writable location
 
-from fastapi import FastAPI, UploadFile, File
+from flask import Flask, request, render_template, jsonify
 from transformers import ViTForImageClassification, ViTFeatureExtractor
 import torch
 import torch.nn as nn
@@ -9,7 +9,7 @@ import torchvision.transforms as transforms
 from PIL import Image
 import io
 
-app = FastAPI()
+app = Flask(__name__)
 
 # Load the ViT model and its feature extractor
 model_name = "google/vit-base-patch16-224-in21k"
@@ -22,7 +22,6 @@ model.classifier = nn.Linear(model.config.hidden_size, num_classes)
 model.load_state_dict(torch.load("skin_cancer_model.pth", map_location=torch.device('cpu')))
 model.eval()
 
-
 # Define class labels
 class_labels = ['benign_keratosis-like_lesions', 'basal_cell_carcinoma', 'actinic_keratoses', 'vascular_lesions', 'melanocytic_Nevi', 'melanoma', 'dermatofibroma']
 
@@ -35,26 +34,60 @@ transform = transforms.Compose([
     transforms.ToTensor(),
 ])
 
-# Define API endpoint for model inference with class-specific thresholds
-@app.post('/predict')
-async def predict(file: UploadFile = File(...)):
-    contents = await file.read()
-    image = Image.open(io.BytesIO(contents))
-    image = transform(image).unsqueeze(0)  # Add batch dimension and move to device
-    
+@app.route('/')
+def index():
+    return render_template('index.html', appName="Skin Cancer Classification Application")
+
+def model_predict(image):
+    image = transform(image).unsqueeze(0)  # Add batch dimension
     with torch.no_grad():
         outputs = model(image)
-    
-    # Calculate softmax probabilities
-    probabilities = torch.softmax(outputs.logits, dim=1).cpu().numpy()[0]
-    
-    # Get predicted class index and its probability
-    predicted_idx = torch.argmax(torch.tensor(probabilities)).item()
-    predicted_label = class_labels[predicted_idx]
-    predicted_probability = probabilities[predicted_idx]
-    
-    # Check if the predicted probability meets the class-specific threshold
-    if predicted_probability < thresholds[predicted_idx]:
-        return {'predicted_class': 'uncertain', 'accuracy': float(predicted_probability)}
+    return outputs
+
+@app.route('/predictApi', methods=["POST"])
+def api():
+    try:
+        if 'fileup' not in request.files:
+            return jsonify({'Error': "Please try again. The Image doesn't exist"})
+        file = request.files.get('fileup')
+        image = Image.open(io.BytesIO(file.read()))
+        result = model_predict(image)
+        
+        probabilities = torch.softmax(result.logits, dim=1).cpu().numpy()[0]
+        predicted_idx = torch.argmax(torch.tensor(probabilities)).item()
+        max_prob = probabilities[predicted_idx]
+        threshold = thresholds[predicted_idx]
+        
+        if max_prob < threshold:
+            return jsonify({'Error': 'No cancer detected or benign lesion.'})
+        prediction = class_labels[predicted_idx]
+        return jsonify({'prediction': prediction})
+    except Exception as e:
+        return jsonify({'Error': 'An error occurred', 'Message': str(e)})
+
+@app.route('/predict', methods=['GET', 'POST'])
+def predict():
+    if request.method == 'POST':
+        try:
+            if 'fileup' not in request.files:
+                return render_template('index.html', prediction='No file selected.', appName="Skin Cancer Classification Application")
+            file = request.files['fileup']
+            image = Image.open(io.BytesIO(file.read()))
+            result = model_predict(image)
+            
+            probabilities = torch.softmax(result.logits, dim=1).cpu().numpy()[0]
+            predicted_idx = torch.argmax(torch.tensor(probabilities)).item()
+            max_prob = probabilities[predicted_idx]
+            threshold = thresholds[predicted_idx]
+            
+            if max_prob < threshold:
+                return render_template('index.html', prediction='No cancer detected or benign lesion.', appName="Skin Cancer Classification Application")
+            prediction = class_labels[predicted_idx]
+            return render_template('index.html', prediction=prediction, appName="Skin Cancer Classification Application")
+        except Exception as e:
+            return render_template('index.html', prediction='Error: ' + str(e), appName="Skin Cancer Classification Application")
     else:
-        return {'predicted_class': predicted_label, 'accuracy': float(predicted_probability)}
+        return render_template('index.html', appName="Skin Cancer Classification Application")
+
+if __name__ == '__main__':
+    app.run(debug=True)