Merge pull request #170 from PEtab-dev/release_0.1.28

Release 0.1.28

Author: dweindl

.github/workflows/ci_tests.yml

@@ -11,7 +11,7 @@ jobs:
     strategy:
       matrix:
         platform: [windows-latest, macos-latest, ubuntu-latest]
-        python-version: ["3.7", "3.10"]
+        python-version: ["3.8", "3.10"]
     runs-on: ${{ matrix.platform }}
 
     steps:

CHANGELOG.md

@@ -2,6 +2,22 @@
 
 ## 0.1 series
 
+### 0.1.28
+
+* Fixed validation for output parameters columns in the condition table
+  by @dweindl in https://github.com/PEtab-dev/libpetab-python/pull/161
+* Added Python support policy
+  by @dweindl in https://github.com/PEtab-dev/libpetab-python/pull/162
+* Fixed typehints and deprecation warning
+  by @dweindl in https://github.com/PEtab-dev/libpetab-python/pull/165
+* Fixed SBML validation
+  by @dweindl in https://github.com/PEtab-dev/libpetab-python/pull/168
+* Fixed deprecation warning from `get_model_for_condition`
+  by @dweindl in https://github.com/PEtab-dev/libpetab-python/pull/169
+
+**Full Changelog**:
+https://github.com/PEtab-dev/libpetab-python/compare/v0.1.27...v0.1.28
+
 ### 0.1.27
 
 Features:

README.md

@@ -22,7 +22,8 @@ and the easiest way to install it is running
 
     pip3 install petab
 
-It will require Python>=3.7.1 to run.
+It will require Python>=3.8 to run. (We are following the
+[numpy Python support policy](https://numpy.org/neps/nep-0029-deprecation_policy.html)).
 
 Development versions of the PEtab library can be installed using
 

petab/conditions.py

@@ -93,9 +93,8 @@ def get_parametric_overrides(condition_df: pd.DataFrame) -> List[str]:
     Returns:
         List of parameter IDs that are mapped in a condition-specific way
     """
-    constant_parameters = list(
-        set(condition_df.columns.values.tolist()) - {CONDITION_ID,
-                                                     CONDITION_NAME})
+    constant_parameters = (set(condition_df.columns.values.tolist())
+                           - {CONDITION_ID, CONDITION_NAME})
     result = []
 
     for column in constant_parameters:
@@ -104,7 +103,5 @@ def get_parametric_overrides(condition_df: pd.DataFrame) -> List[str]:
 
         floatified = condition_df.loc[:, column].apply(core.to_float_if_float)
 
-        for x in floatified:
-            if not isinstance(x, float):
-                result.append(x)
+        result.extend(x for x in floatified if not isinstance(x, float))
     return result

petab/core.py

@@ -20,7 +20,7 @@
            'create_combine_archive', 'unique_preserve_order']
 
 
-def get_simulation_df(simulation_file: str) -> pd.DataFrame:
+def get_simulation_df(simulation_file: Union[str, Path]) -> pd.DataFrame:
     """Read PEtab simulation table
 
     Arguments:
@@ -33,7 +33,7 @@ def get_simulation_df(simulation_file: str) -> pd.DataFrame:
                        float_precision='round_trip')
 
 
-def write_simulation_df(df: pd.DataFrame, filename: str) -> None:
+def write_simulation_df(df: pd.DataFrame, filename: Union[str, Path]) -> None:
     """Write PEtab simulation table
 
     Arguments:
@@ -91,7 +91,8 @@ def get_notnull_columns(df: pd.DataFrame, candidates: Iterable):
 
 
 def flatten_timepoint_specific_output_overrides(
-        petab_problem: 'petab.problem.Problem') -> None:
+        petab_problem: 'petab.problem.Problem',
+) -> None:
     """Flatten timepoint-specific output parameter overrides.
 
     If the PEtab problem definition has timepoint-specific

petab/lint.py

@@ -60,9 +60,7 @@ def _check_df(df: pd.DataFrame, req_cols: Iterable, name: str) -> None:
     Raises:
           AssertionError: if a column is missing
     """
-    cols_set = df.columns.values
-    missing_cols = set(req_cols) - set(cols_set)
-    if missing_cols:
+    if missing_cols := set(req_cols) - set(df.columns.values):
         raise AssertionError(
             f"DataFrame {name} requires the columns {missing_cols}.")
 
@@ -85,12 +83,16 @@ def assert_no_leading_trailing_whitespace(
 
 
 def check_condition_df(
-        df: pd.DataFrame, model: Optional[Model] = None) -> None:
+        df: pd.DataFrame,
+        model: Optional[Model] = None,
+        observable_df: Optional[pd.DataFrame] = None
+) -> None:
     """Run sanity checks on PEtab condition table
 
     Arguments:
         df: PEtab condition DataFrame
         model: Model for additional checking of parameter IDs
+        observable_df: PEtab observables DataFrame
 
     Raises:
         AssertionError: in case of problems
@@ -101,7 +103,7 @@ def check_condition_df(
     _check_df(df, req_cols, "condition")
 
     # Check for correct index
-    if not df.index.name == CONDITION_ID:
+    if df.index.name != CONDITION_ID:
         raise AssertionError(
             f"Condition table has wrong index {df.index.name}."
             f"expected {CONDITION_ID}.")
@@ -119,6 +121,9 @@ def check_condition_df(
 
     if model is not None:
         allowed_cols = set(model.get_valid_ids_for_condition_table())
+        if observable_df is not None:
+            allowed_cols |= set(petab.get_output_parameters(
+                model=model, observable_df=observable_df))
         for column_name in df.columns:
             if column_name != CONDITION_NAME \
                     and column_name not in allowed_cols:
@@ -154,14 +159,9 @@ def check_measurement_df(df: pd.DataFrame,
                 df[column_name].values, column_name)
 
     if observable_df is not None:
-        # Check all observables are defined
-        observables_defined = set(observable_df.index.values)
-        observables_used = set(df[OBSERVABLE_ID])
-        observables_undefined = observables_used - observables_defined
-        if observables_undefined:
-            raise ValueError(f"Observables {observables_undefined} used in "
-                             "measurement table but not defined in "
-                             "observables table.")
+        assert_measured_observables_defined(df, observable_df)
+        measurements.assert_overrides_match_parameter_count(
+            df, observable_df)
 
         if OBSERVABLE_TRANSFORMATION in observable_df:
             # Check for positivity of measurements in case of
@@ -176,11 +176,6 @@ def check_measurement_df(df: pd.DataFrame,
                                      f'transformation {trafo} must be '
                                      f'positive, but {measurement} <= 0.')
 
-    if observable_df is not None:
-        assert_measured_observables_defined(df, observable_df)
-        measurements.assert_overrides_match_parameter_count(
-            df, observable_df)
-
     assert_measurements_not_null(df)
     assert_measurements_numeric(df)
 
@@ -206,7 +201,7 @@ def check_parameter_df(
 
     _check_df(df, PARAMETER_DF_REQUIRED_COLS[1:], "parameter")
 
-    if not df.index.name == PARAMETER_ID:
+    if df.index.name != PARAMETER_ID:
         raise AssertionError(
             f"Parameter table has wrong index {df.index.name}."
             f"expected {PARAMETER_ID}.")
@@ -232,10 +227,11 @@ def check_parameter_df(
                                  f"but column {NOMINAL_VALUE} is missing.")
         try:
             df.loc[non_estimated_par_ids, NOMINAL_VALUE].apply(float)
-        except ValueError:
-            raise AssertionError("Expected numeric values for "
-                                 f"`{NOMINAL_VALUE}` in parameter table for "
-                                 "all non-estimated parameters.")
+        except ValueError as e:
+            raise AssertionError(
+                f"Expected numeric values for `{NOMINAL_VALUE}` in parameter "
+                "table for all non-estimated parameters."
+            ) from e
 
     assert_parameter_id_is_string(df)
     assert_parameter_scale_is_valid(df)
@@ -285,8 +281,9 @@ def check_observable_df(observable_df: pd.DataFrame) -> None:
         try:
             sp.sympify(obs)
         except sp.SympifyError as e:
-            raise AssertionError(f"Cannot parse expression '{obs}' "
-                                 f"for observable {row.Index}: {e}")
+            raise AssertionError(
+                f"Cannot parse expression '{obs}' "
+                f"for observable {row.Index}: {e}") from e
 
         noise = getattr(row, NOISE_FORMULA)
         try:
@@ -297,9 +294,10 @@ def check_observable_df(observable_df: pd.DataFrame) -> None:
                 raise AssertionError(f"No or non-finite {NOISE_FORMULA} "
                                      f"given for observable {row.Index}.")
         except sp.SympifyError as e:
-            raise AssertionError(f"Cannot parse expression '{noise}' "
-                                 f"for noise model for observable "
-                                 f"{row.Index}: {e}")
+            raise AssertionError(
+                f"Cannot parse expression '{noise}' "
+                f"for noise model for observable " f"{row.Index}: {e}"
+            ) from e
 
 
 def assert_all_parameters_present_in_parameter_df(
@@ -346,7 +344,8 @@ def assert_all_parameters_present_in_parameter_df(
 
 def assert_measured_observables_defined(
         measurement_df: pd.DataFrame,
-        observable_df: pd.DataFrame) -> None:
+        observable_df: pd.DataFrame
+) -> None:
     """Check if all observables in the measurement table have been defined in
     the observable table
 
@@ -360,12 +359,11 @@ def assert_measured_observables_defined(
 
     used_observables = set(measurement_df[OBSERVABLE_ID].values)
     defined_observables = set(observable_df.index.values)
-    undefined_observables = used_observables - defined_observables
-
-    if undefined_observables:
+    if undefined_observables := (used_observables - defined_observables):
         raise AssertionError(
-            "Undefined observables in measurement file: "
-            f"{undefined_observables}.")
+            f"Observables {undefined_observables} used in "
+            "measurement table but not defined in observables table."
+        )
 
 
 def condition_table_is_parameter_free(condition_df: pd.DataFrame) -> bool:
@@ -540,9 +538,10 @@ def assert_parameter_prior_parameters_are_valid(
                 pars = tuple(
                     float(val) for val in pars_str.split(PARAMETER_SEPARATOR)
                 )
-            except ValueError:
+            except ValueError as e:
                 raise AssertionError(
-                    f"Could not parse prior parameters '{pars_str}'.")
+                    f"Could not parse prior parameters '{pars_str}'.") from e
+
             # all distributions take 2 parameters
             if len(pars) != 2:
                 raise AssertionError(
@@ -795,7 +794,8 @@ def lint_problem(problem: 'petab.Problem') -> bool:
     if problem.condition_df is not None:
         logger.info("Checking condition table...")
         try:
-            check_condition_df(problem.condition_df, problem.model)
+            check_condition_df(problem.condition_df, problem.model,
+                               problem.observable_df)
         except AssertionError as e:
             logger.error(e)
             errors_occurred = True
@@ -925,9 +925,7 @@ def assert_measurement_conditions_present_in_condition_table(
         used_conditions |= \
             set(measurement_df[PREEQUILIBRATION_CONDITION_ID].dropna().values)
     available_conditions = set(condition_df.index.values)
-    missing_conditions = used_conditions - available_conditions
-
-    if missing_conditions:
+    if missing_conditions := (used_conditions - available_conditions):
         raise AssertionError("Measurement table references conditions that "
                              "are not specified in the condition table: "
                              + str(missing_conditions))

petab/measurements.py

@@ -2,6 +2,7 @@
 # noqa: F405
 
 import itertools
+import math
 import numbers
 from pathlib import Path
 from typing import Dict, List, Union
@@ -201,7 +202,7 @@ def create_measurement_df() -> pd.DataFrame:
         Created DataFrame
     """
 
-    df = pd.DataFrame(data={
+    return pd.DataFrame(data={
         OBSERVABLE_ID: [],
         PREEQUILIBRATION_CONDITION_ID: [],
         SIMULATION_CONDITION_ID: [],
@@ -213,8 +214,6 @@ def create_measurement_df() -> pd.DataFrame:
         REPLICATE_ID: []
     })
 
-    return df
-
 
 def measurements_have_replicates(measurement_df: pd.DataFrame) -> bool:
     """Tests whether the measurements come with replicates
@@ -235,15 +234,15 @@ def measurements_have_replicates(measurement_df: pd.DataFrame) -> bool:
 
 def assert_overrides_match_parameter_count(
         measurement_df: pd.DataFrame,
-        observable_df: pd.DataFrame) -> None:
+        observable_df: pd.DataFrame
+) -> None:
     """Ensure that number of parameters in the observable definition matches
     the number of overrides in ``measurement_df``
 
     Arguments:
         measurement_df: PEtab measurement table
         observable_df: PEtab observable table
     """
-
     # sympify only once and save number of parameters
     observable_parameters_count = {
         obs_id: len(observables.get_formula_placeholders(
@@ -260,10 +259,11 @@ def assert_overrides_match_parameter_count(
         # check observable parameters
         try:
             expected = observable_parameters_count[row[OBSERVABLE_ID]]
-        except KeyError:
+        except KeyError as e:
             raise ValueError(
                 f"Observable {row[OBSERVABLE_ID]} used in measurement table "
-                f"is not defined.")
+                f"is not defined.") from e
+
         actual = len(split_parameter_replacement_list(
             row.get(OBSERVABLE_PARAMETERS, None)))
         # No overrides are also allowed
@@ -289,7 +289,7 @@ def assert_overrides_match_parameter_count(
         except KeyError:
             # no overrides defined, but a numerical sigma can be provided
             # anyways
-            if not len(replacements) == 1 \
+            if len(replacements) != 1 \
                     or not isinstance(replacements[0], numbers.Number):
                 raise AssertionError(
                     f'No placeholders have been specified in the noise model '

petab/models/sbml_model.py

@@ -9,7 +9,7 @@
 from . import MODEL_TYPE_SBML
 from .model import Model
 from ..sbml import (get_sbml_model, is_sbml_consistent, load_sbml_from_string,
-                    log_sbml_errors, write_sbml)
+                    write_sbml)
 
 
 class SbmlModel(Model):
@@ -109,9 +109,7 @@ def symbol_allowed_in_observable_formula(self, id_: str) -> bool:
         return self.sbml_model.getElementBySId(id_) or id_ == 'time'
 
     def is_valid(self) -> bool:
-        valid = is_sbml_consistent(self.sbml_model.getSBMLDocument())
-        log_sbml_errors(self.sbml_model.getSBMLDocument())
-        return valid
+        return is_sbml_consistent(self.sbml_model.getSBMLDocument())
 
     def is_state_variable(self, id_: str) -> bool:
         return (self.sbml_model.getSpecies(id_) is not None