remove knn

app.py

@@ -191,18 +191,14 @@ def load_sample(data, current_index):
     image_id = data[current_index]["id"]
     qimage = data[current_index]["image"]
     
-    neighbors_path = os.path.join(knn_cache_path, f"{image_id}.JPEG")
-    neighbors_image = Image.open(neighbors_path).convert('RGB')
-    
     labels = data[current_index]["correct_label"]
-    return qimage, neighbors_image, labels
+    return qimage, labels
     # return qimage, neighbors_image, training_samples_image
 
 
 def update_app(decision, data, current_index, history, username):
     if current_index == -1:
         data = generate_dataset()
-        nns = {}
     
     if current_index>=0 and current_index < NUMBER_OF_IMAGES-1:
         time_stamp = int(time.time())
@@ -233,20 +229,19 @@ def update_app(decision, data, current_index, history, username):
         os.remove(temp_filename)
 
     elif current_index == NUMBER_OF_IMAGES-1:
-        return None, None, None, current_index, history, data, None, None
+        return None, None, current_index, history, data, None
 
     current_index += 1
-    qimage, neighbors_image, labels = load_sample(data, current_index)
+    qimage, labels = load_sample(data, current_index)
     image_id = data[current_index]["id"]
     training_samples_image = get_training_samples(image_id)
     training_samples_image = [Image.open(x).convert('RGB') for x in training_samples_image]
-    nns = label_dist_of_nns(image_id)
 
     # labels is a list of labels, conver it to a string
     labels = ", ".join(labels)
     label_plot = string_to_image(labels)
 
-    return qimage, label_plot, neighbors_image, current_index, history, data, nns, training_samples_image
+    return qimage, label_plot, current_index, history, data, training_samples_image
 
 
 newcss = '''
@@ -286,26 +281,26 @@ with gr.Blocks(css=newcss) as demo:
             with gr.Column():
                 label_plot = gr.Plot(label='Is this a correct label for this image?', type='fig')
                 training_samples = gr.Gallery(type="pil", label="Training samples" , elem_id="sample_gallery")
-    with gr.Column():
-        gr.Markdown("## Nearest Neighbors Analysis of the Query (ResNet-50)")
-        nn_labels = gr.Label(label="NN-Labels")
-        neighbors_image = gr.Image(type="pil", label="Nearest Neighbors", elem_id="nn_gallery")
+    # with gr.Column():
+    #     gr.Markdown("## Nearest Neighbors Analysis of the Query (ResNet-50)")
+    #     nn_labels = gr.Label(label="NN-Labels")
+    #     neighbors_image = gr.Image(type="pil", label="Nearest Neighbors", elem_id="nn_gallery")
 
     accept_btn.click(
         update_app,
         inputs=[accept_btn, data_gr, current_index, history, username],
-        outputs=[query_image, label_plot, neighbors_image, current_index, history, data_gr, nn_labels, training_samples]
+        outputs=[query_image, label_plot, current_index, history, data_gr, training_samples]
     )
     myabe_btn.click(
         update_app,
         inputs=[myabe_btn, data_gr, current_index, history, username],
-        outputs=[query_image, label_plot, neighbors_image, current_index, history, data_gr, nn_labels, training_samples]
+        outputs=[query_image, label_plot, current_index, history, data_gr, training_samples]
     )
 
     reject_btn.click(
         update_app,
         inputs=[reject_btn, data_gr, current_index, history, username],
-        outputs=[query_image, label_plot, neighbors_image, current_index, history, data_gr, nn_labels, training_samples]
+        outputs=[query_image, label_plot, current_index, history, data_gr, training_samples]
     )
 
 demo.launch()