Use pandas.testing.assert_frame_equal and skimage.metrics.structural_…

…similarity

for comparison when testing

conftest.py

@@ -1,11 +1,5 @@
-import hashlib
-import shutil
-from pathlib import Path
-
 import pytest
 
-from isimip_qa.main import init_settings
-
 
 def pytest_addoption(parser):
     parser.addoption('--init-files', dest='init_files', action='store_true', default=False)
@@ -17,33 +11,3 @@ class Config:
         init_files = pytestconfig.getoption('init_files')
 
     return Config
-
-
-@pytest.fixture(scope='session')
-def settings(config):
-    if config.init_files:
-        extractions_path = Path('testing') / 'extractions'
-        plots_path = Path('testing') / 'plots'
-    else:
-        extractions_path = plots_path = Path('testing') / 'tmp'
-
-    shutil.rmtree(extractions_path, ignore_errors=True)
-    shutil.rmtree(plots_path, ignore_errors=True)
-
-    return init_settings(
-        datasets_path=Path('testing') / 'datasets',
-        extractions_path=extractions_path,
-        plots_path=plots_path,
-        plots_format='png'
-    )
-
-
-@pytest.fixture()
-def checksum():
-    def sha1_checksum(file_path):
-        h = hashlib.sha1()
-        with open(file_path, 'rb') as fp:
-            h.update(fp.read())
-        return h.hexdigest()
-
-    return sha1_checksum

isimip_qa/tests/test_extractions.py

@@ -1,8 +1,11 @@
+import shutil
 from pathlib import Path
 
+import pandas as pd
 import pytest
 
 from isimip_qa.extractions import extraction_classes
+from isimip_qa.main import init_settings
 from isimip_qa.models import Dataset
 
 datasets = [
@@ -12,9 +15,20 @@
     'sine'
 ]
 
+@pytest.fixture(scope='module')
+def settings(config):
+    extractions_path = Path('testing') / ('extractions' if config.init_files else 'tmp')
+
+    shutil.rmtree(extractions_path, ignore_errors=True)
+
+    return init_settings(
+        datasets_path=Path('testing') / 'datasets',
+        extractions_path=extractions_path
+    )
+
 @pytest.mark.parametrize('dataset_path', datasets)
 @pytest.mark.parametrize('extraction_class', extraction_classes)
-def test_extraction(config, settings, checksum, dataset_path, extraction_class):
+def test_extraction(settings, dataset_path, extraction_class):
     dataset = Dataset(dataset_path)
     extraction = extraction_class(dataset)
     extraction.path.unlink(missing_ok=True)
@@ -25,9 +39,15 @@ def test_extraction(config, settings, checksum, dataset_path, extraction_class):
         file.close()
 
     assert extraction.path.exists()
+    assert extraction.path.suffix in ['.csv', '.json']
+
+    template_path = Path('testing').joinpath('extractions') \
+                                   .joinpath(extraction.path.relative_to(settings.EXTRACTIONS_PATH))
 
-    if not config.init_files:
-        assert checksum(extraction.path) == checksum(
-            Path('testing').joinpath('extractions')
-                           .joinpath(extraction.path.relative_to(settings.EXTRACTIONS_PATH))
+    if extraction.path.suffix == '.csv':
+        pd.testing.assert_frame_equal(
+            pd.read_csv(extraction.path),
+            pd.read_csv(template_path)
         )
+    else:
+        assert extraction.path.read_text() == template_path.read_text()

isimip_qa/tests/test_plots.py

@@ -1,8 +1,12 @@
+import shutil
 from pathlib import Path
 
 import pytest
+from skimage.io import imread
+from skimage.metrics import structural_similarity as ssim
 
 from isimip_qa.extractions import extraction_classes
+from isimip_qa.main import init_settings
 from isimip_qa.models import Dataset
 from isimip_qa.plots import plot_classes
 
@@ -13,10 +17,24 @@
     'sine'
 ]
 
+@pytest.fixture(scope='module')
+def settings(config):
+    extractions_path = Path('testing') / 'extractions'
+    plots_path = Path('testing') / ('plots' if config.init_files else 'tmp')
+
+    shutil.rmtree(plots_path, ignore_errors=True)
+
+    return init_settings(
+        datasets_path=Path('testing') / 'datasets',
+        extractions_path=extractions_path,
+        plots_path=plots_path,
+        plots_format='png'
+    )
+
 @pytest.mark.parametrize('dataset_path', datasets)
 @pytest.mark.parametrize('extraction_class', extraction_classes)
 @pytest.mark.parametrize('plot_class', plot_classes)
-def test_mean_extraction(config, settings, checksum, dataset_path, extraction_class, plot_class):
+def test_plot(settings, dataset_path, extraction_class, plot_class):
     dataset = Dataset(dataset_path)
     if plot_class.has_extraction(extraction_class):
         plot = plot_class(extraction_class, [dataset], path=dataset.path.stem)
@@ -25,8 +43,12 @@ def test_mean_extraction(config, settings, checksum, dataset_path, extraction_cl
         for sp in plot.get_subplots():
             assert sp.path.exists()
 
-            if not config.init_files:
-                assert checksum(sp.path) == checksum(
-                    Path('testing').joinpath('plots')
-                                   .joinpath(sp.path.relative_to(settings.PLOTS_PATH))
-                )
+            img = imread(sp.path, as_gray=True)
+
+            template_path = Path('testing').joinpath('plots') \
+                                           .joinpath(sp.path.relative_to(settings.PLOTS_PATH))
+            template_img = imread(template_path, as_gray=True)
+
+            similarity = ssim(img, template_img, data_range=template_img.max() - template_img.min())
+
+            assert similarity == 1.0

pyproject.toml

@@ -45,6 +45,7 @@ dev = [
     "pytest",
     "pytest-cov",
     "ruff",
+    "scikit-image",
     "twine"
 ]