Three.py

import numpy as np
from sklearn.neighbors import NearestNeighbors

# Convert the feature vectors to a NumPy array
features_array = np.array(list(features_dict.values()))

# Create a NearestNeighbors object and fit it to the features array
knn = NearestNeighbors(n_neighbors=11, metric='cosine')
knn.fit(features_array)

# Define a function to retrieve the most similar images to a query image
def retrieve_similar_images(query_image_name, knn_model, features_dict):
    # Get the features for the query image
    query_features = features_dict[query_image_name]
    # Reshape the features to match the expected input format for the knn_model
    query_features = query_features.reshape(1, -