Improved header-data alignment checks

.github/workflows/ci.yml

@@ -20,6 +20,8 @@ jobs:
         uses: actions/checkout@v4
       - name: Set up Python
         uses: actions/setup-python@v5
+        with:
+          python-version: '3.11'
       - name: Setup Pixi
         uses: prefix-dev/setup-pixi@v0.5.1
       - name: Run linter

geost/base.py

@@ -93,7 +93,7 @@ def __new__(cls, *args, **kwargs):
     def __repr__(self):
         return f"{self.__class__.__name__}:\n# header = {self.n_points}"
 
-    def reset_header(self):
+    def reset_header(self, *args):
         """
         Create a new header based on the 'data' dataframe
         (:py:attr:`~geost.base.PointDataCollection.data`). Can be used to reset the
@@ -200,9 +200,8 @@ def header(self, header):
         header through validation and warn the user of any potential problems.
         """
         headerschema.validate(header)
-        if any(~header["nr"].isin(self.data["nr"].unique())):
-            warn("Header does not cover all unique objects in data")
         self._header = header
+        self.__check_header_to_data_alignment()
 
     @data.setter
     def data(self, data):
@@ -221,6 +220,34 @@ def data(self, data):
             inclined_dataschema.validate(data)
 
         self._data = data
+        self.__check_header_to_data_alignment()
+
+    def __check_header_to_data_alignment(self):
+        """
+        Two-way check to warn of any misalignment between the header and data
+        attributes. Two way, i.e. if header includes more objects than in the data and
+        if the data includes more unique objects that listed in the header.
+
+        This check is performed everytime the object is instantiated AND if any change
+        is made to either the header or data attributes (see their respective setters).
+        """
+        if hasattr(self, "_header") and hasattr(self, "_data"):
+            if any(~self.header["nr"].isin(self.data["nr"].unique())):
+                warn(
+                    "Header covers more objects than present in the data table, "
+                    "consider running the method 'reset_header' to update the header."
+                )
+            if any(
+                [
+                    True
+                    for nr in self.data["nr"].unique()
+                    if not self.header["nr"].isin([nr]).any()
+                ]
+            ):
+                warn(
+                    "Header does not cover all unique objects in data, consider running "
+                    + "the method 'reset_header' to update the header."
+                )
 
     def change_vertical_reference(self, to: str):
         """
@@ -730,7 +757,9 @@ def slice_depth_interval(
 
         return result
 
-    def slice_by_values(self, column: str, selection_values: Union[str, Iterable]):
+    def slice_by_values(
+        self, column: str, selection_values: Union[str, Iterable], invert: bool = False
+    ):
         """
         Slice rows from data based on matching condition. E.g. only return rows with
         a certain lithology in the collection object.
@@ -742,6 +771,9 @@ def slice_by_values(self, column: str, selection_values: Union[str, Iterable]):
             values.
         selection_values : Union[str, Iterable]
             Values to look for in the column.
+        invert : bool
+            Invert the slicing action, so remove layers with selected values instead of
+            keeping them.
 
         Returns
         -------
@@ -754,7 +786,11 @@ def slice_by_values(self, column: str, selection_values: Union[str, Iterable]):
             selection_values = [selection_values]
 
         data_sliced = self.data.copy()
-        data_sliced = data_sliced[data_sliced[column].isin(selection_values)]
+        if invert:
+            data_sliced = data_sliced[~data_sliced[column].isin(selection_values)]
+        elif not invert:
+            data_sliced = data_sliced[data_sliced[column].isin(selection_values)]
+
         header_sliced = self.header.loc[
             self.header["nr"].isin(data_sliced["nr"].unique())
         ]

tests/test_export.py

@@ -112,7 +112,7 @@ def test_vtk_prepare_borehole(self, borehole_collection):
         )
         assert_array_equal(prepared_borehole, target)
 
-    @pytest.mark.skip(reason="CI known for not working with pyvisyta")
+    @pytest.mark.unittest  # .skip(reason="Gitlab CI not working with pyvisyta")
     def test_vtk_borehole_to_multiblock(self, borehole_collection):
         multiblock = vtk.borehole_to_multiblock(
             borehole_collection.data,
@@ -126,7 +126,7 @@ def test_vtk_borehole_to_multiblock(self, borehole_collection):
         assert multiblock[0].n_cells == 22
         assert multiblock[0].n_points == 260
 
-    @pytest.mark.skip(reason="CI known for not working with pyvisyta")
+    @pytest.mark.unittest  # .skip(reason="Gitlab CI not working with pyvisyta")
     def test_to_vtm(self, borehole_collection):
         out_file = self.export_folder.joinpath("test_output_file.vtm")
         out_folder = self.export_folder.joinpath("test_output_file")

tests/test_pointcollection.py

@@ -98,6 +98,20 @@ def header_missing_object(self):
 
         return dataframe
 
+    @pytest.fixture
+    def header_surplus_objects(self):
+        dataframe = pd.DataFrame(
+            {
+                "nr": ["B-01", "B-02", "B-03"],
+                "x": [139370, 100000, 110000],
+                "y": [455540, 400000, 410000],
+                "mv": [1, 0, -1],
+                "end": [-4, -8, -9],
+            }
+        )
+
+        return dataframe
+
     @pytest.mark.unittest
     def test_change_vertical_reference(self, borehole_df_ok):
         borehole_collection_ok = BoreholeCollection(borehole_df_ok)
@@ -241,7 +255,15 @@ def test_validation_fail(self, capfd, borehole_df_bad_validation):
         assert 'data in column "end" failed check "< mv" for 1 rows: [1]' in out
 
     @pytest.mark.integrationtest
-    def test_header_mismatch(self, capfd, borehole_df_ok, header_missing_object):
-        BoreholeCollection(borehole_df_ok, header=header_missing_object)
+    def test_header_mismatch(
+        self, capfd, borehole_df_ok, header_missing_object, header_surplus_objects
+    ):
+        # Situation #1: More unique objects in data table than listed in header
+        collection = BoreholeCollection(borehole_df_ok, header=header_missing_object)
         out, err = capfd.readouterr()
         assert "Header does not cover all unique objects in data" in out
+
+        # Situation #2: More objects in header table than in data table
+        collection = BoreholeCollection(borehole_df_ok, header=header_surplus_objects)
+        out, err = capfd.readouterr()
+        assert "Header covers more objects than present in the data table" in out