Update src/app.py

src/app.py

@@ -49,12 +49,6 @@ def check_index(ds):
 
     else:
         return index_dataset(ds)
-
-
-def find_similar_images(method):
-    if method == "FAISS":
-        
-    return retrieved examples
     
 
 dataset_with_embeddings = check_index(candidate_subset)
@@ -62,7 +56,7 @@ dataset_with_embeddings = check_index(candidate_subset)
 # Main function, to find similar images
 # TODO: implement different distance measures
 
-def get_neighbors(query_image, selected_descriptor, top_k=5):
+def get_neighbors(query_image, selected_descriptor, selected_distance top_k=5):
     """Returns the top k nearest examples to the query image.
 
     Args:
@@ -76,16 +70,15 @@ def get_neighbors(query_image, selected_descriptor, top_k=5):
         cd = ColorDescriptor((8, 12, 3))
         qi_embedding = cd.describe(query_image)
         qi_np = np.array(qi_embedding)
-        if distance_measure == "FAISS":
+        if selected_distance == "FAISS":
             scores, retrieved_examples = dataset_with_embeddings.get_nearest_examples(
                 'color_embeddings', qi_np, k=top_k)
-        elif distance_measure == "Chi-squared":
+        elif selected_distance == "Chi-squared":
             tmp_dataset = dataset_with_embeddings.map(lambda row: {'distance': chi2_distance(histA=query_vector, histB=row['color_embeddings'])})
             retrieved_examples = tmp_dataset.sort("distance")
         else: 
             tmp_dataset = dataset_with_embeddings.map(lambda row: {'distance': euclidian_distance(histA=query_vector, histB=row['color_embeddings'])})
             retrieved_examples = tmp_dataset.sort("distance")
-  
         images = retrieved_examples['image'] #retrieved images is a dict, with images and embeddings
         return images
     if "CLIP" == selected_descriptor:
@@ -133,13 +126,13 @@ with gr.Blocks() as demo:
 
     with gr.Row():
         descr_dropdown = gr.Dropdown(["Color Descriptor", "LBP", "CLIP"], value="LBP", label="Please choose an descriptor")
-        checkboxes_descr = gr.CheckboxGroup(["Color Descriptor", "LBP", "CLIP"], label="Please choose an descriptor")
+        dist_dropdown = gr.Dropdown(["FAISS", "Chi-squared", "Euclid"], value="FAISS", label="Please choose a distance measure")
         dataset_dropdown = gr.Dropdown(
         ["huggan/CelebA-faces", "EIT/cbir-eit"],
         value="huggan/CelebA-faces",
         label="Please select a dataset"
     )
-    btn.click(get_neighbors, inputs=[image_input, descr_dropdown], outputs=[gallery_output])
+    btn.click(get_neighbors, inputs=[image_input, descr_dropdown, dist_dropdown], outputs=[gallery_output])
 
 
 if __name__ == "__main__":