Create test_meters.py

Test file for the meters.py file

geotils/ToBeChecked/metrics/test_meters.py

@@ -0,0 +1,103 @@
+import unittest
+
+class TestMeters(unittest.TestCase):
+    def test_reset(self):
+        #testing if the 2 meters are actually resetting
+        avg_meter = AverageMeter()
+        avg_meter.value = 5.0
+        avg_meter.average = 3.0
+        avg_meter.count = 10
+        avg_meter.reset()
+        self.assertEqual(avg_meter.count, 0)
+        self.assertEqual(avg_meter.value, 0.0)
+        self.assertEqual(avg_meter.average, 0.0)
+
+        short_term_meter = ShortTermMemoryMeter(memory_length=3)
+        short_term_meter.value = 5.0
+        short_term_meter.length = 3
+        short_term_meter.in_memory = [3.0, 4.0, 5.0]
+        short_term_meter.average = 4.0
+        short_term_meter.reset()
+        self.assertEqual(short_term_meter.length, 0)
+        self.assertEqual(short_term_meter.value, 0)
+        self.assertEqual(short_term_meter.average, 0)
+
+    def test_update(self):
+        avg_meter = AverageMeter()
+        avg_meter.value = 4.0
+        avg_meter.count = 1
+        avg_meter.average = 4.0
+        avg_meter.update(5.0)
+        self.assertEqual(avg_meter.count, 2)
+        self.assertEqual(avg_meter.value, 5.0)
+        self.assertEqual(avg_meter.average, 4.5)
+
+        short_term_meter = ShortTermMemoryMeter(memory_length=3)
+        short_term_meter.update(2.0)
+        self.assertEqual(short_term_meter.length, 1)
+        self.assertEqual(short_term_meter.average, 2.0)
+
+class TestSampler(unittest.TestCase):
+    def test_reset(self):
+        sampler = Sampler(value_names=['value1', 'value2'], rate=2)
+        sampler.reset()
+        self.assertEqual(sampler.history, {'value1': [], 'value2': []})
+
+    def test_update(self):
+        sampler = Sampler(value_names=['value1', 'value2'], rate=2)
+        sampler.update([1, 2], 0)
+        self.assertEqual(sampler.history, {'value1': [1], 'value2': [2]})
+
+class DummyCriterion(torch.nn.Module):
+    def __init__(self):
+        super(DummyCriterion, self).__init__()
+        self.__name__="Module"
+
+    def forward(self, x, y):
+        return x + y
+
+class DummyMetric:
+    def __init__(self):
+        self.__name__=""
+
+    def get_update(self):
+        return 0.5
+
+class TestMetricsManager(unittest.TestCase):
+    def test_initialization(self):
+        classes = ['class1', 'class2']
+        metric_names = ['iou', 'fscore']
+        thresh = 0.5
+
+        manager = MetricsManager(classes, metric_names, thresh)
+
+        self.assertEqual(manager.classes, classes)
+        self.assertEqual(manager.metric_names, metric_names)
+        self.assertEqual(manager.thresh, thresh)
+        self.assertEqual(len(manager.metrics), len(classes) * len(metric_names))
+
+    def test_register_scores(self):
+        classes = ['class1', 'class2']
+        metric_names = ['iou', 'fscore']
+        thresh = 0.5
+
+        manager = MetricsManager(classes, metric_names, thresh)
+
+        pdict = {}
+        seg_loss_meter = AverageMeter()
+        seg_criterion = DummyCriterion()
+        cls_loss_meter = AverageMeter()
+        cls_criterion = DummyCriterion()
+        cls_accuracy_meter = AverageMeter()
+        cls_f1score_meter = AverageMeter()
+        dm1 = DummyMetric()
+        dm1.__name__="metric1"
+        dm2 = DummyMetric()
+        dm2.__name__="metric2"
+        mnms = [(dm1, AverageMeter()), (dm2, AverageMeter())]
+
+        pdict = manager.register_scores(pdict, seg_loss_meter, seg_criterion, cls_loss_meter, cls_criterion, mnms, cls_accuracy_meter, cls_f1score_meter)
+
+        expected_length = len(classes) * len(metric_names) + 2
+        self.assertEqual(len(pdict), expected_length)
+