Allow label transform functions

src/cr/cube/dimension.py

@@ -14,7 +14,7 @@
     MARGINAL,
     MEASURE,
 )
-from cr.cube.util import lazyproperty
+from cr.cube.util import common_prefix, common_suffix, lazyproperty
 
 from .util import format, format_datetime
 
@@ -33,7 +33,71 @@
 }
 
 
-def _formatter(dimension_type, typedef, out_format) -> Union[Callable, partial]:
+class LabelTransformFuncs:
+    # This is the system missing. We shouldn't consider it for the labels
+    # when looking for common prefix. This isn't a user-set value.
+    SYSTEM_LABELS = {"No Data"}
+
+    def __init__(self, label_transforms: list[dict], all_labels: list[str]):
+        """
+        Used to apply the label function transformations on elements of a
+        dimension
+
+        :param label_transforms: list of dictionary functions. See
+        Dashboard-models for the list of options and arguments
+        :param all_labels: list of all the original labels of the dimension. It
+        is necessary to have them all so we can do the prefix/suffix trimming.
+        """
+        self.label_transforms = label_transforms
+        self.all_labels = [l for l in all_labels if l not in self.SYSTEM_LABELS]
+
+    def apply(self, formatter: Union[Callable, partial]) -> Callable[[str], str]:
+        """
+        Receives the formatter and will wrap any other transformation
+        on top of its output.
+        :param formatter: Callable that receives a sting and outputs a string
+        :return: A callable with the same signature as the formatter
+        """
+        if not self.label_transforms:
+            return formatter
+
+        def __inner(value: str) -> str:
+            for func_transform in self.label_transforms:
+                # Invalid function raises AttributeError, this should break
+                # because it's an incomplete implementation. The functions
+                # should be specified in the Lark syntax and Dashboard models
+                # to match the list of allowed operations here. Requires
+                # 3 repos to be updated.
+                func = getattr(self, func_transform["function"])
+                args = func_transform["args"]
+                value = func(value, args)
+
+            value = formatter(value)
+            return value
+
+        return __inner
+
+    @staticmethod
+    def replace(value: str, args) -> str:
+        return value.replace(*args)
+
+    def remove_common_suffix(self, value: str, args) -> str:
+        suffix_pos = common_suffix(self.all_labels)
+        return value[:suffix_pos] if suffix_pos else value
+
+    def remove_common_prefix(self, value: str, args) -> str:
+        prefix_pos = common_prefix(self.all_labels)
+        return value[prefix_pos:] if prefix_pos else prefix_pos
+
+    def trim_common(self, value: str, args) -> str:
+        no_suffix = self.remove_common_suffix(value, [])
+        no_prefix_either = self.remove_common_prefix(no_suffix, [])
+        return no_prefix_either
+
+
+def _formatter(
+    dimension_type, typedef, out_format, label_transforms, all_labels
+) -> Union[Callable, partial]:
     """Returns a formatting function according to the dimension type."""
 
     if dimension_type != DT.DATETIME:
@@ -46,6 +110,10 @@ def _formatter(dimension_type, typedef, out_format) -> Union[Callable, partial]:
             if orig_format and out_format
             else format
         )
+
+    if label_transforms:  # Apply label transforms on the output
+        formatter = LabelTransformFuncs(label_transforms, all_labels).apply(formatter)
+
     return formatter
 
 
@@ -534,6 +602,7 @@ def from_typedef(
             element_defs = [codemap[code] for code in order if code in codemap]
 
         all_xforms = dimension_transforms_dict.get("elements", {})
+        label_transforms = dimension_transforms_dict.get("label_transforms")
         if dimension_type == DT.MR_SUBVAR:
             hidden_xforms = cls._hidden_transforms(
                 element_defs,
@@ -542,13 +611,20 @@ def from_typedef(
             all_xforms = {**hidden_xforms, **all_xforms}
 
         elements = []
+        all_labels = [elt["name"] for elt in element_defs if "name" in elt]
         for idx, element_dict in enumerate(element_defs):
             # --- convert to string for categorical ids
             element_id = _build_element_id(element_dict, dimension_type)
             xforms = _ElementTransforms(
                 all_xforms.get(element_id, all_xforms.get(str(element_id), {}))
             )
-            formatter = _formatter(dimension_type, typedef, element_data_format)
+            formatter = _formatter(
+                dimension_type,
+                typedef,
+                element_data_format,
+                label_transforms,
+                all_labels,
+            )
             element = Element(element_dict, idx, xforms, formatter, dimension_type)
             elements.append(element)
 
@@ -934,7 +1010,7 @@ def __init__(
         self._element_dict = element_dict
         self._index = index
         self._element_transforms = element_transforms
-        self._label_formatter = label_formatter
+        self._label_formatter: Callable = label_formatter
         self._dim_type = dim_type
 
     def __repr__(self) -> str:
@@ -1005,7 +1081,10 @@ def label(self) -> str:
         This value is the empty string when no value has been specified or display of
         the name has been suppressed.
         """
-        return self._str_representation_for("name")
+        _label = self._str_representation_for("name")
+        if self._label_formatter is not None:
+            _label = self._label_formatter(_label)
+        return _label
 
     @lazyproperty
     def missing(self) -> bool:

src/cr/cube/util.py

@@ -110,3 +110,22 @@ def __set__(self, obj, value):
         quite snappy and probably not a rich target for optimization efforts.
         """
         raise AttributeError("can't set attribute")
+
+
+def common_prefix(str_list):
+    prefix_pos = 0
+    for pos, char in enumerate(str_list[0]):
+        try:
+            if {s[pos] for s in str_list[1:]} != {char}:
+                break
+            prefix_pos += 1
+        except IndexError:
+            # This means the pivot is now beyond the shortest word, this
+            # means that word is the common prefix
+            return prefix_pos
+    return prefix_pos
+
+
+def common_suffix(str_list):
+    str_list = [list(reversed(s)) for s in str_list]
+    return -1 * common_prefix(str_list)  # ;)

tests/integration/test_cubepart.py

@@ -1224,6 +1224,21 @@ def test_it_can_sort_rows_by_column_percent(self):
         actual = np.round(slice_.column_percentages, 1).tolist()
         assert expected == actual, "\n%s\n\n%s" % (expected, actual)
 
+    def test_it_applies_label_transforms(self):
+        transforms = {
+            "rows_dimension": {
+                "label_transforms": [
+                    {"function": "replace", "args": ["nough", "XXX"]},
+                ]
+            }
+        }
+        slice_ = _Slice(Cube(CR.CAT_4_X_CAT_5), 0, transforms, None, 0)
+
+        # Note the labels have been replaced from "Enough" to "eXXX"
+        expected = ['Plenty', 'EXXX', 'Not eXXX', 'N/A']
+        actual = slice_.row_labels.tolist()
+        assert expected == actual, "\n%s\n\n%s" % (expected, actual)
+
     def test_it_can_sort_rows_by_labels(self):
         """Responds to order:label sort-by-label."""
         transforms = {

tests/unit/test_dimension.py

@@ -6,10 +6,11 @@
 import pytest
 
 from cr.cube.dimension import (
-    Element,
-    Elements,
     Dimension,
     Dimensions,
+    Element,
+    Elements,
+    LabelTransformFuncs,
     _ElementIdShim,
     _ElementTransforms,
     _OrderSpec,
@@ -23,13 +24,7 @@
     MEASURE,
 )
 
-from ..unitutil import (
-    call,
-    class_mock,
-    instance_mock,
-    method_mock,
-    property_mock,
-)
+from ..unitutil import call, class_mock, instance_mock, method_mock, property_mock
 
 
 class TestDimensions:
@@ -1190,6 +1185,25 @@ def test_it_knows_its_label(
         element = Element(element_dict, None, element_transforms_, str, None)
         assert element.label == expected_value
 
+    def test_label_function(self):
+        typedef = {
+            "class": "categorical",
+            "categories": [
+                {"id": 1, "name": "xx 1. REPLaced"},
+                {"id": 2, "name": "xx 2. REPLace me"},
+            ],
+        }
+        transforms_dict = {
+            "label_transforms": [
+                {"function": "replace", "args": ["REPL", "repl"]},
+                {"function": "trim_common", "args": []},
+                {"function": "replace", "args": ["me", "you"]},
+            ],
+        }
+        elements = Elements.from_typedef(typedef, transforms_dict, DT.CAT, None)
+        assert elements[0].label == "1. replaced"
+        assert elements[1].label == "2. replace you"
+
     @pytest.mark.parametrize(
         ("hide", "expected_value"), ((True, True), (False, False), (None, False))
     )
@@ -1773,3 +1787,36 @@ def subtrahend_ids_(self, request):
     @pytest.fixture
     def valid_elements_(self, request):
         return instance_mock(request, Elements)
+
+
+class TestLabelTransformFuncs:
+    def test_prefix(self):
+        all_labels = ["a1", "a2", "a3"]
+        transformer = LabelTransformFuncs([], all_labels)
+        assert transformer.remove_common_prefix("a2", []) == "2"
+
+    def test_suffix(self):
+        all_labels = ["a1b", "a2b", "a3b"]
+        transformer = LabelTransformFuncs([], all_labels)
+        assert transformer.remove_common_suffix("a2b", []) == "a2"
+
+    def test_trim_common(self):
+        all_labels = ["a1b", "a2b", "a3b"]
+        transformer = LabelTransformFuncs([], all_labels)
+        assert transformer.trim_common("a2b", []) == "2"
+
+    def test_replace(self):
+        all_labels = ["a1b", "a2b", "a3b"]
+        transformer = LabelTransformFuncs([], all_labels)
+        assert transformer.replace("a2b", ["a", "AA"]) == "AA2b"
+
+    def test_application(self):
+        all_labels = ["a1b", "a2b", "a3b"]
+        transforms = [
+            {"function": "trim_common", "args": []},
+            {"function": "replace", "args": ["1", "one"]},
+            {"function": "replace", "args": ["2", "two"]},
+        ]
+        transformer = LabelTransformFuncs(transforms, all_labels)
+        formatter = transformer.apply(lambda x: x.upper())
+        assert formatter("a2b") == "TWO"