Move to python3 plus update dependencies

Author: varun-suresh

clustering.py

@@ -93,8 +93,7 @@ def calculate_symmetric_dist(app_nearest_neighbors):
     for row_no, row_val in enumerate(results):
         d[row_no, :] = row_val
     d_time = time()-dist_calc_time
-    print 'Distance calculation time : {}'.format(d_time)
-
+    print("Distance calculation time : {}".format(d_time))
     return d
 
 
@@ -184,7 +183,7 @@ def cluster(descriptor_matrix, n_neighbors=10, thresh=[2]):
     clusters = []
     for th in thresh:
         clusters_th = aro_clustering(app_nearest_neighbors, distance_matrix, th)
-        print 'N Clusters: {}, thresh: {}'.format(len(clusters_th), th)
+        print("N Clusters: {}, thresh: {}".format(len(clusters_th), th))
         clusters.append({'clusters': clusters_th, 'threshold': th})
     return clusters
 

demo.py

@@ -18,10 +18,10 @@ def plot_histogram(lfw_dir):
         for dirname in dirs:
             n_photos = len(os.listdir(os.path.join(root, dirname)))
             filecount_dict[dirname] = n_photos
-    print 'No of unique people: {}'.format(len(filecount_dict.keys()))
+    print("No of unique people: {}".format(len(filecount_dict.keys())))
     df = pd.DataFrame(filecount_dict.items(), columns=['Name', 'Count'])
-    print 'Singletons : {}\nTwo :{}\n'.format((df['Count'] == 1).sum(),
-                                              (df['Count'] == 2).sum())
+    print("Singletons : {}\nTwo :{}\n".format((df['Count'] == 1).sum(),
+                                              (df['Count'] == 2).sum()))
     plt.hist(df['Count'], bins=max(df['Count']))
     plt.title('Cluster Sizes')
     plt.xlabel('No of images in folder')
@@ -63,11 +63,11 @@ def evaluate_clusters(clusters, labels_lookup):
     """
     precision, recall = calculate_pairwise_pr(clusters, labels_lookup)
     f1_score = 2*precision*recall/(precision+recall)
-    print 'Precision : {}\nRecall : {}\nf1_score : {}'.format(precision,
+    print("Precision : {}\nRecall : {}\nf1_score : {}".format(precision,
                                                               recall,
                                                               f1_score
-                                                              )
-    print '---------------------------------------------------------'
+                                                              ))
+    print("---------------------------------------------------------")
     return f1_score
 
 
@@ -108,8 +108,8 @@ def create_labels_lookup(labels):
 
         labels_lookup = create_labels_lookup(labels)
         for clusters in clusters_thresholds:
-            print 'No of clusters: {}'.format(len(clusters['clusters']))
-            print 'Threshold : {}'.format(clusters['threshold'])
+            print("No of clusters: {}".format(len(clusters['clusters'])))
+            print("Threshold : {}".format(clusters['threshold']))
             f1_score = evaluate_clusters(clusters['clusters'], labels_lookup)
         # n_faces = 0
         # for c in clusters:

evaluation.py

@@ -33,15 +33,15 @@ def calculate_pairwise_pr(clusters, labels_lookup):
     # Recall:
     gt_clusters = defaultdict(list)
     # Count the actual number of possible true pairs:
-    for row_no, label in labels_lookup.iteritems():
+    for row_no, label in labels_lookup.items():
         gt_clusters[label].append(row_no)
     true_pairs = 0
-    for cluster_id, cluster_items in gt_clusters.iteritems():
+    for cluster_id, cluster_items in gt_clusters.items():
         n = len(cluster_items)
         true_pairs += n * (n-1)/2.0
-    print 'Correct Pairs that are in the same cluster:{}'.format(correct_pairs)
-    print 'Total pairs as per the clusters created: {}'.format(total_pairs)
-    print 'Total possible true pairs:{}'.format(true_pairs)
+    print("Correct Pairs that are in the same cluster:{}".format(correct_pairs))
+    print("Total pairs as per the clusters created: {}".format(total_pairs))
+    print("Total possible true pairs:{}".format(true_pairs))
     precision = float(correct_pairs)/total_pairs
     recall = float(correct_pairs)/true_pairs
     return precision, recall

setup.py

@@ -5,17 +5,28 @@
     find_packages)
 
 runtime_packages = [
-    'flask == 0.12.3',
-    'numpy==1.12.1',
-    'matplotlib==1.5.2',
-    'profilehooks==1.9.0',
-    'pandas==0.20.2',
-    'scipy==0.19.0'
+"Click==7.0",
+"cycler==0.10.0",
+"Flask==1.1.1",
+"itsdangerous==1.1.0",
+"Jinja2==2.10.1",
+"kiwisolver==1.1.0",
+"MarkupSafe==1.1.1",
+"matplotlib==3.1.1",
+"numpy==1.17.0",
+"pandas==0.25.0",
+"profilehooks==1.11.0",
+"pyparsing==2.4.1.1",
+"python-dateutil==2.8.0",
+"pytz==2019.1",
+"scipy==1.3.0",
+"six==1.12.0",
+"Werkzeug==0.15.5",
 ]
 
 setup(
     name='approximate-nn-clustering',
-    version='0.0.1',
+    version='0.0.2',
     description='A modified version of rank order clustering',
     author='Varun Suresh',
     author_email='[email protected]',

visualize.py

@@ -32,7 +32,7 @@ def create_idx_to_path(lfw_dir):
 
 @app.route('/img/<path:fpath>', methods=['GET'])
 def get_img_path(fpath):
-    print os.path.dirname(fpath), fpath.split('/')[-1]
+    print(os.path.dirname(fpath), fpath.split('/')[-1])
     return send_from_directory(os.path.dirname(fpath),fpath.split('/')[-1])
 
 @app.route('/single_cluster', methods=["GET"])