Column.get(map=) support dicts with partial keys (#403)

audeering · Nov 6, 2023 · 69332d4 · 69332d4
1 parent 5152318
commit 69332d4
Show file tree

Hide file tree

Showing 3 changed files with 28 additions and 10 deletions.
diff --git a/audformat/core/column.py b/audformat/core/column.py
@@ -254,20 +254,30 @@ def get(
                     "for its labels."
                 )
 
+            # Check that at least one key is available for map
+            # if labels are stored as dictionary
+            keys = []
+            for key, value in labels.items():
+                if isinstance(value, dict):
+                    keys += list(value.keys())
+            keys = sorted(list(set(keys)))
+            if len(keys) > 0 and map not in keys:
+                raise ValueError(
+                    f"Cannot map "
+                    f"'{self._id}' "
+                    f"to "
+                    f"'{map}'. "
+                    f"Expected one of "
+                    f"{list(keys)}."
+                )
+
             mapping = {}
             for key, value in labels.items():
                 if isinstance(value, dict):
                     if map in value:
                         value = value[map]
                     else:
-                        raise ValueError(
-                            f"Cannot map "
-                            f"'{self._id}' "
-                            f"to "
-                            f"'{map}'. "
-                            f"Expected one of "
-                            f"{list(value)}."
-                        )
+                        value = np.NaN
                 mapping[key] = value
 
             result = result.map(mapping)

diff --git a/tests/conftest.py b/tests/conftest.py
@@ -8,6 +8,11 @@
 
 
 pytest.DB = audformat.testing.create_db()
+# Adjust scheme dictionary to contain one missing label
+labels = pytest.DB.schemes["label_map_str"].labels
+labels['label3'].pop('prop2')
+pytest.DB.schemes["label_map_str"].replace_labels(labels)
+
 pytest.DB_ROOT = 'db'
 pytest.FILE_DUR = pd.to_timedelta('1s')
 

diff --git a/tests/test_column.py b/tests/test_column.py
@@ -177,7 +177,7 @@ def test_get_as_segmented():
             pytest.DB['files']['string'], 'map', None,
             marks=pytest.mark.xfail(raises=ValueError),
         ),
-        pytest.param(  # no labels
+        pytest.param(  # no labels in dict
             pytest.DB['files']['label_map_str'], 'bad', None,
             marks=pytest.mark.xfail(raises=ValueError),
         ),
@@ -189,7 +189,10 @@ def test_map(column, map, expected_dtype):
     mapping = {}
     for key, value in pytest.DB.schemes[column.scheme_id].labels.items():
         if isinstance(value, dict):
-            value = value[map]
+            if map in value:
+                value = value[map]
+            else:
+                value = np.NaN
         mapping[key] = value
     expected = expected.map(mapping).astype(expected_dtype)
     expected.name = map