Ensure measurements are numeric and not null with linter (#76)

Co-authored-by: Daniel Weindl <[email protected]>

Author: dilpath

petab/lint.py

@@ -20,6 +20,8 @@
 __all__ = ['assert_all_parameters_present_in_parameter_df',
            'assert_measured_observables_defined',
            'assert_measurement_conditions_present_in_condition_table',
+           'assert_measurements_not_null',
+           'assert_measurements_numeric',
            'assert_model_parameters_in_condition_or_parameter_table',
            'assert_no_leading_trailing_whitespace',
            'assert_noise_distributions_valid',
@@ -177,6 +179,9 @@ def check_measurement_df(df: pd.DataFrame,
         measurements.assert_overrides_match_parameter_count(
             df, observable_df)
 
+    assert_measurements_not_null(df)
+    assert_measurements_numeric(df)
+
 
 def check_parameter_df(
         df: pd.DataFrame,
@@ -906,6 +911,51 @@ def assert_measurement_conditions_present_in_condition_table(
                              + str(missing_conditions))
 
 
+def assert_measurements_not_null(
+    measurement_df: pd.DataFrame,
+) -> None:
+    """Check whether all measurements are not null.
+
+    Arguments:
+        measurement_df:
+            PEtab measurement table.
+
+    Raises:
+        AssertionError:
+            Some measurement value(s) are null (missing).
+    """
+    if measurement_df[MEASUREMENT].isnull().any():
+        raise AssertionError('Some measurement(s) are null (missing).')
+
+
+def assert_measurements_numeric(
+    measurement_df: pd.DataFrame,
+) -> None:
+    """Check whether all measurements are numeric.
+
+    Note that null (missing) measurements are ignored.
+
+    Arguments:
+        measurement_df:
+            PEtab measurement table.
+
+    Raises:
+        AssertionError:
+            Some measurement value(s) are not numeric.
+    """
+    not_null_measurement_values = measurement_df[MEASUREMENT].dropna()
+    all_measurements_are_numeric = (
+        pd.to_numeric(not_null_measurement_values, errors='coerce')
+        .notnull()
+        .all()
+    )
+    if not all_measurements_are_numeric:
+        raise AssertionError(
+            'Some values in the `petab.C.MEASUREMENT` column of the PEtab '
+            'measurements table are not numeric.'
+        )
+
+
 def is_valid_identifier(x: str) -> bool:
     """Check whether `x` is a valid identifier
 

tests/test_petab.py

@@ -340,7 +340,7 @@ def test_flatten_timepoint_specific_output_overrides():
         TIME:
             [1.0, 1.0, 2.0, 2.0],
         MEASUREMENT:
-            [np.nan] * 4,
+            [.1] * 4,
         OBSERVABLE_PARAMETERS:
             ['obsParOverride1;1.0', 'obsParOverride2;1.0',
              'obsParOverride2;1.0', 'obsParOverride2;1.0'],
@@ -362,7 +362,7 @@ def test_flatten_timepoint_specific_output_overrides():
         TIME:
             [1.0, 1.0, 2.0, 2.0],
         MEASUREMENT:
-            [np.nan] * 4,
+            [.1] * 4,
         OBSERVABLE_PARAMETERS:
             ['obsParOverride1;1.0', 'obsParOverride2;1.0',
              'obsParOverride2;1.0', 'obsParOverride2;1.0'],
@@ -425,7 +425,7 @@ def test_flatten_timepoint_specific_output_overrides_special_cases():
         TIME:
             [1.0, 1.0, 2.0, 2.0],
         MEASUREMENT:
-            [np.nan] * 4,
+            [.1] * 4,
         NOISE_PARAMETERS:
             ['noiseParOverride1', 'noiseParOverride1',
              'noiseParOverride2', 'noiseParOverride2'],
@@ -442,7 +442,7 @@ def test_flatten_timepoint_specific_output_overrides_special_cases():
         TIME:
             [1.0, 1.0, 2.0, 2.0],
         MEASUREMENT:
-            [np.nan] * 4,
+            [.1] * 4,
         NOISE_PARAMETERS:
             ['noiseParOverride1', 'noiseParOverride1',
              'noiseParOverride2', 'noiseParOverride2'],