Merge pull request #12 from LSSTDESC/issue/9/consolidate-e2e-script

Remove e2e tests script, and replace it with equivalent unit tests

src/resspect/scripts/fit_dataset.py

@@ -69,20 +69,20 @@ def fit_dataset(user_choices):
     features_file = user_choices.output
     ncores = user_choices.ncores
 
-    if user_choices.sim_name in ['SNPCC', 'snpcc']:
+    if user_choices.sim_name.lower() == 'snpcc':
         # fit the entire sample
         fit_snpcc(path_to_data_dir=data_dir, features_file=features_file,
                   number_of_processors=ncores, feature_extractor=user_choices.function)
 
-    elif user_choices.sim_name in ['PLAsTiCC', 'PLASTICC', 'plasticc']:
+    elif user_choices.sim_name.lower() == 'plasticc':
         fit_plasticc(path_photo_file=user_choices.photo_file,
                     path_header_file=user_choices.header_file,
                     output_file=features_file,
                     sample=user_choices.sample,
                     number_of_processors=ncores)
 
-    return None
-
+    else:
+        raise ValueError("-s or --simulation not recognized. Options are 'SNPCC' or 'PLAsTiCC'.")
 
 def main():
 

src/resspect/scripts/run_loop.py

@@ -15,7 +15,7 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-__all__ = ['learn_loop', 'run_loop']
+__all__ = ['run_loop']
 
 from resspect.learn_loop import learn_loop
 
@@ -63,23 +63,31 @@ def run_loop(args):
 
     """
 
-    # set training sample variable
-    if args.training == 'original':
-        train = 'original'
-    elif isinstance(int(args.training), int):
-        train = int(args.training)
-    else:
-        raise ValueError('-t or --training option must be '
-                         '"original" or integer!')
-
     # run active learning loop
-    learn_loop(nloops=args.nquery, features_method=args.method,
+    learn_loop(nloops=args.nquery,
+               features_method=args.method,
                classifier=args.classifier,
-               strategy=args.strategy, path_to_features=args.input,
+               strategy=args.strategy,
+               path_to_features=args.input,
                output_metrics_file=args.metrics,
                output_queried_file=args.queried,
-               training=train, batch=args.batch)
+               training=_parse_training(args.training),
+               batch=args.batch)
+
+def _parse_training(training:str):
+    """We don't check that `isinstance(training, str)` because `training` is defined
+    as a string in the argparse.ArgumentParser.
+    """
+    # set training sample variable
+    if training.lower() == 'original':
+        train = 'original'
+    else:
+        try:
+            train = int(training)
+        except ValueError:
+            raise ValueError('-t or --training option must be "original" or integer!')
 
+    return train
 
 def main():
 

tests/resspect/test_learn_loop.py

@@ -24,5 +24,30 @@ def test_can_run_learn_loop(test_des_data_path):
             output_queried_file=os.path.join(dir_name,"just_other_name.csv"),
         )
 
+
+def test_can_run_learn_loop_uncsample(test_des_data_path):
+    """Test that learn_loop can load data and run.
+    This instance is distinct from the previous because it uses `UncSample` strategy
+    and runs for 2 loops instead of 1.
+    """
+    with tempfile.TemporaryDirectory() as dir_name:
+        # Create the feature files to use for the learning loop.
+        output_file = os.path.join(dir_name, "output_file.dat")
+        fit_snpcc(
+            path_to_data_dir=test_des_data_path,
+            features_file=output_file
+        )
+
+        learn_loop(
+            nloops=2,
+            features_method="bazin",
+            strategy="UncSampling",
+            path_to_features=output_file,
+            output_metrics_file=os.path.join(dir_name,"just_a_name.csv"),
+            output_queried_file=os.path.join(dir_name,"just_other_name.csv"),
+            training=10,
+        )
+
+
 if __name__ == '__main__':
     pytest.main()

tests/test_cli_scripts.py

@@ -0,0 +1,20 @@
+import pytest
+
+
+def test_run_loop_arg_check():
+    """Test that the logic that parses CLI inputs works as expected. Specifically
+    that the training sample is correctly parsed as either an integer or the string 'original'.
+    """
+
+    from resspect.scripts.run_loop import _parse_training
+
+    assert _parse_training('original') == 'original'
+    assert _parse_training('OrigINAL') == 'original'
+    assert _parse_training('10') == 10
+    with pytest.raises(ValueError):
+        _parse_training('not_a_number')
+    with pytest.raises(ValueError):
+        _parse_training('1.0')
+    with pytest.raises(ValueError):
+        _parse_training('')
+    assert _parse_training('010') == 10