Plot apd

examples/sample_notebooks/evidently_metrics.ipynb

@@ -45,6 +45,7 @@
     "from evidently.metrics import ColumnDriftMetric\n",
     "from evidently.metrics import DataDriftTable\n",
     "from evidently.metrics import DatasetDriftMetric\n",
+    "from evidently.metrics import ColumnCategoryMetric\n",
     "from evidently.metrics import ColumnDistributionMetric\n",
     "from evidently.metrics import ColumnValuePlot\n",
     "from evidently.metrics import ColumnQuantileMetric\n",
@@ -305,6 +306,7 @@
     "    ColumnCorrelationsMetric(column_name=\"education\"),\n",
     "    ColumnValueListMetric(column_name=\"relationship\", values=[\"Husband\", \"Unmarried\"]), \n",
     "    ColumnValueRangeMetric(column_name=\"age\", left=10, right=20),\n",
+    "    ColumnCategoryMetric(column_name='education', category='Some-college'),\n",
     "    \n",
     "])\n",
     "\n",

examples/sample_notebooks/evidently_tests.ipynb

@@ -225,7 +225,9 @@
     "    TestColumnNumberOfDifferentMissingValues(column_name='education'),\n",
     "    TestColumnAllConstantValues(column_name='education'),\n",
     "    TestColumnAllUniqueValues(column_name='education'),\n",
-    "    TestColumnRegExp(column_name='education',reg_exp='^[0..9]')\n",
+    "    TestColumnRegExp(column_name='education',reg_exp='^[0..9]'),\n",
+    "    TestCategoryShare(column_name='education', category='Some-college', lt=0.5),\n",
+    "    TestCategoryShare(column_name='age', category=27., lt=0.5)\n",
     "])\n",
     "\n",
     "data_integrity_column_tests.run(reference_data=adult_ref, current_data=adult_cur)\n",

src/evidently/calculations/data_drift.py

@@ -210,7 +210,7 @@ def get_one_column_drift(
                 column_name,
                 datetime_column_name,
             )
-            current_scatter["current"] = df
+            current_scatter["current (mean)"] = df
             if prefix is None:
                 x_name = "Index binned"
             else:

src/evidently/metrics/__init__.py

@@ -29,6 +29,7 @@
 from .data_integrity.column_summary_metric import ColumnSummaryMetric
 from .data_integrity.dataset_missing_values_metric import DatasetMissingValuesMetric
 from .data_integrity.dataset_summary_metric import DatasetSummaryMetric
+from .data_quality.column_category_metric import ColumnCategoryMetric
 from .data_quality.column_correlations_metric import ColumnCorrelationsMetric
 from .data_quality.column_distribution_metric import ColumnDistributionMetric
 from .data_quality.column_quantile_metric import ColumnQuantileMetric
@@ -76,6 +77,7 @@
     "ColumnSummaryMetric",
     "DatasetMissingValuesMetric",
     "DatasetSummaryMetric",
+    "ColumnCategoryMetric",
     "ColumnCorrelationsMetric",
     "ColumnDistributionMetric",
     "ColumnInteractionPlot",

src/evidently/metrics/data_drift/column_drift_metric.py

@@ -149,7 +149,7 @@ def get_one_column_drift(
                 column.name,
                 datetime_name,
             )
-            current_scatter["current"] = df
+            current_scatter["current (mean)"] = df
             if prefix is None:
                 x_name = "Index binned"
             else:
@@ -372,9 +372,10 @@ def render_html(self, obj: ColumnDriftMetric) -> List[BaseWidgetInfo]:
                     std=(result.scatter.plot_shape["y0"] - result.scatter.plot_shape["y1"]) / 2,
                     xaxis_name=result.scatter.x_name,
                     xaxis_name_ref=None,
-                    yaxis_name=result.column_name,
+                    yaxis_name=f"{result.column_name} (mean +/- std)",
                     color_options=self.color_options,
                     return_json=False,
+                    line_name="reference (mean)",
                 )
             tabs.append(TabData("DATA DRIFT", plotly_figure(title="", figure=scatter_fig)))
 

src/evidently/metrics/data_drift/data_drift_table.py

@@ -201,9 +201,10 @@ def _generate_column_params(
                         std=(data.scatter.plot_shape["y0"] - data.scatter.plot_shape["y1"]) / 2,
                         xaxis_name=data.scatter.x_name,
                         xaxis_name_ref=None,
-                        yaxis_name=data.column_name,
+                        yaxis_name=f"{data.column_name} (mean +/- std)",
                         color_options=self.color_options,
                         return_json=False,
+                        line_name="reference (mean)",
                     )
                 scatter = plotly_figure(title="", figure=scatter_fig)
                 details.with_part("DATA DRIFT", info=scatter)

src/evidently/metrics/data_drift/text_descriptors_drift_metric.py

@@ -173,9 +173,10 @@ def _generate_column_params(
                     std=(data.scatter.plot_shape["y0"] - data.scatter.plot_shape["y1"]) / 2,
                     xaxis_name=data.scatter.x_name,
                     xaxis_name_ref=None,
-                    yaxis_name=data.column_name,
+                    yaxis_name=f"{data.column_name} (mean +/- std)",
                     color_options=self.color_options,
                     return_json=False,
+                    line_name="reference (mean)",
                 )
             scatter = plotly_figure(title="", figure=scatter_fig)
             details.with_part("DATA DRIFT", info=scatter)

src/evidently/metrics/data_quality/column_category_metric.py

@@ -0,0 +1,111 @@
+from typing import Dict
+from typing import List
+from typing import Optional
+from typing import Union
+
+import pandas as pd
+
+from evidently.base_metric import ColumnName
+from evidently.base_metric import InputData
+from evidently.base_metric import Metric
+from evidently.base_metric import MetricResult
+from evidently.model.widget import BaseWidgetInfo
+from evidently.options.base import AnyOptions
+from evidently.renderers.base_renderer import MetricRenderer
+from evidently.renderers.base_renderer import default_renderer
+from evidently.renderers.html_widgets import CounterData
+from evidently.renderers.html_widgets import counter
+from evidently.renderers.html_widgets import header_text
+
+
+class CategoryStat(MetricResult):
+    all_num: int
+    category_num: int
+    category_ratio: float
+
+
+class ColumnCategoryMetricResult(MetricResult):
+    column_name: str
+    category: Union[int, float, str]
+    current: CategoryStat
+    reference: Optional[CategoryStat] = None
+    counts_of_values: Dict[str, pd.DataFrame]
+
+
+class ColumnCategoryMetric(Metric[ColumnCategoryMetricResult]):
+    """Calculates count and shares of values in the predefined values list"""
+
+    column_name: ColumnName
+    category: Union[int, float, str]
+
+    def __init__(
+        self, column_name: Union[str, ColumnName], category: Union[int, float, str], options: AnyOptions = None
+    ) -> None:
+        self.column_name = ColumnName.from_any(column_name)
+        self.category = category
+        super().__init__(options=options)
+
+    def calculate(self, data: InputData) -> ColumnCategoryMetricResult:
+        if not data.has_column(self.column_name):
+            raise ValueError(f"Column '{self.column_name.display_name}' was not found in data.")
+
+        current_column = data.get_current_column(self.column_name)
+        reference_column = data.get_reference_column(self.column_name)
+
+        counts_of_values = None
+        counts_of_values = {}
+        current_counts = current_column.value_counts(dropna=False).reset_index()
+        current_counts.columns = ["x", "count"]
+        counts_of_values["current"] = current_counts.head(10)
+        counts_of_values["current"].index = counts_of_values["current"].index.astype("str")
+        if reference_column is not None:
+            reference_counts = reference_column.value_counts(dropna=False).reset_index()
+            reference_counts.columns = ["x", "count"]
+            counts_of_values["reference"] = reference_counts.head(10)
+            counts_of_values["reference"].index = counts_of_values["reference"].index.astype("str")
+
+        reference: Optional[CategoryStat] = None
+        if reference_column is not None:
+            reference = CategoryStat(
+                all_num=len(reference_column),
+                category_num=(reference_column == self.category).sum(),
+                category_ratio=(reference_column == self.category).mean(),
+            )
+        return ColumnCategoryMetricResult(
+            column_name=self.column_name.display_name,
+            category=self.category,
+            current=CategoryStat(
+                all_num=current_column.shape[0],
+                category_num=(current_column == self.category).sum(),
+                category_ratio=(current_column == self.category).mean(),
+            ),
+            reference=reference,
+            counts_of_values=counts_of_values,
+        )
+
+
+@default_renderer(wrap_type=ColumnCategoryMetric)
+class ColumnCategoryMetricRenderer(MetricRenderer):
+    def _get_count_info(self, stat: CategoryStat):
+        percents = round(stat.category_ratio * 100, 3)
+        return f"{stat.category_num} out of {stat.all_num} ({percents}%)"
+
+    def render_html(self, obj: ColumnCategoryMetric) -> List[BaseWidgetInfo]:
+        metric_result = obj.get_result()
+        result = [header_text(label=f"Column '{metric_result.column_name}'. Сategory '{metric_result.category}'.")]
+        counters = [
+            CounterData.string(
+                label="current",
+                value=self._get_count_info(metric_result.current),
+            ),
+        ]
+
+        if metric_result.reference is not None:
+            counters.append(
+                CounterData.string(
+                    label="reference",
+                    value=self._get_count_info(metric_result.reference),
+                ),
+            )
+        result.append(counter(counters=counters))
+        return result

src/evidently/tests/__init__.py

@@ -46,6 +46,8 @@
 from .data_integrity_tests import TestShareOfRowsWithMissingValues
 from .data_quality_tests import TestAllColumnsMostCommonValueShare
 from .data_quality_tests import TestCatColumnsOutOfListValues
+from .data_quality_tests import TestCategoryCount
+from .data_quality_tests import TestCategoryShare
 from .data_quality_tests import TestColumnQuantile
 from .data_quality_tests import TestColumnValueMax
 from .data_quality_tests import TestColumnValueMean
@@ -93,6 +95,8 @@
     "TestTNR",
     "TestTPR",
     "TestAllFeaturesValueDrift",
+    "TestCategoryCount",
+    "TestCategoryShare",
     "TestColumnDrift",
     "TestCustomFeaturesValueDrift",
     "TestEmbeddingsDrift",