resolve sklearn update issues

README.md

@@ -57,6 +57,7 @@ This takes an argument `num_class` which is default to `100`
 
 The following function are possible with the `Recognition` class:
 
+
 Name | Description
 :--- | :---
 update(`targets, predictions`)|takes in a (batch of) vector predictions and their corresponding groundtruth. vector size must match `num_class` in the class initialization.
@@ -138,7 +139,9 @@ This takes an argument `num_class` which is default to `100` and `num_tool` whic
 The following function are possible with the `Detection` class:
 
 Name | Description
+
 :--- | :---
+
 update(`targets, predictions, format`)|input: takes in a (batch of) list/dict predictions and their corresponding groundtruth. Each frame prediction/groundtruth can be either as a `list of list` or `list of dict`. (more details below).
 video_end()|Call to make the end of one video sequence.
 reset()|Reset current records. Useful during training and can be called at the begining of each epoch to avoid overlapping epoch performances.

ivtmetrics/recognition.py

@@ -58,6 +58,10 @@ def __init__(self, num_class=100, ignore_null=False):
         self.num_class = num_class
         self.ignore_null = ignore_null
         self.reset_global()   
+
+    def resolve_nan(self, classwise):
+        classwise[classwise==-0.0] = np.nan
+        return classwise
 
     ##%%%%%%%%%%%%%%%%%%% RESET OP #%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
     def reset(self):
@@ -114,6 +118,7 @@ def compute_AP(self, component="ivt", ignore_null=False):
         with warnings.catch_warnings():
             warnings.filterwarnings(action='ignore') #, message='[info] triplet classes not represented in this test sample will be reported as nan values.')            
             classwise = average_precision_score(targets, predicts, average=None)
+            classwise = self.resolve_nan(classwise)
             if (ignore_null and component=="ivt"): classwise = classwise[:-6]
             mean      = np.nanmean(classwise)
         return {"AP":classwise, "mAP":mean}
@@ -147,6 +152,7 @@ def compute_global_AP(self, component="ivt", ignore_null=False):
         with warnings.catch_warnings():
             warnings.filterwarnings(action='ignore') #, message='[info] triplet classes not represented in this test sample will be reported as nan values.')            
             classwise = average_precision_score(targets, predicts, average=None)
+            classwise = self.resolve_nan(classwise)
             if (ignore_null and component=="ivt"): classwise = classwise[:-6]
             mean      = np.nanmean(classwise)
         return {"AP":classwise, "mAP":mean}    
@@ -180,6 +186,7 @@ def compute_video_AP(self, component="ivt", ignore_null=False):
                 else:
                     sys.exit("Function filtering {} not yet supported!".format(component))                        
                 classwise = average_precision_score(targets, predicts, average=None)
+                classwise = self.resolve_nan(classwise)
                 video_log.append( classwise.reshape([1,-1]) )
             video_log = np.concatenate(video_log, axis=0)         
             videowise = np.nanmean(video_log, axis=0)
@@ -250,6 +257,7 @@ def topClass(self, k=10, component="ivt"):
         else:
             sys.exit("Function filtering {} not supported yet!".format(component))            
         classwise = average_precision_score(targets, predicts, average=None)
+        classwise = self.resolve_nan(classwise)
         pd_idx    = (-classwise).argsort()[:k]
         output    = {x:classwise[x] for x in pd_idx}
         return output

setup.py

@@ -6,7 +6,7 @@
 
 setup(
     name='ivtmetrics',
-    version='0.1.0',    
+    version='0.1.1',    
     packages=['ivtmetrics'],
     author='Chinedu Nwoye',
     author_email='nwoye.chinedu@gmail.com',    
@@ -15,10 +15,10 @@
     long_description = long_description,
     long_description_content_type ='text/x-rst',
     url='https://github.com/CAMMA-public/ivtmetrics',
-    download_url = 'https://github.com/CAMMA-public/ivtmetrics/archive/refs/tags/v0.1.0.tar.gz', 
+    download_url = 'https://github.com/CAMMA-public/ivtmetrics/archive/refs/tags/v0.1.1.tar.gz', 
     include_package_data=True,
     license='BSD 2-clause', # Chose a license from here: https://help.github.com/articles/licensing-a-repository
-    install_requires=['scikit-learn',
+    install_requires=['scikit-learn>=1.0.2',
                       'numpy>=1.21',
                       ],