RowTests for TextEvals (#1691)

* RowTests for TextEvals
Fix row tests charts

* fix lint

* fix tests

Author: mike0sv

src/evidently/metrics/__init__.py

@@ -66,9 +66,11 @@
 from .regression import DummyRMSE
 from .regression import MeanError
 from .regression import R2Score
+from .row_test_summary import RowTestSummary
 
 __all__ = [
     "GroupBy",
+    "RowTestSummary",
     # column statistics metrics
     "CategoryCount",
     "ValueDrift",

src/evidently/metrics/row_test_summary.py

@@ -45,9 +45,9 @@ def _render_test_column_widget(
         assert isinstance(result, SingleValue)
         success_count = int(result.value * row_count)
         if result.value == 1:
-            return pie_chart(title=title, data=(["PASSED"], [success_count]), colors=["GREEN"], size=size)
+            return pie_chart(title=title, data=(["PASSED"], [row_count]), colors=["GREEN"], size=size)
         if result.value == 0:
-            return pie_chart(title=title, data=(["FAILED"], [0]), colors=["RED"], size=size)
+            return pie_chart(title=title, data=(["FAILED"], [row_count]), colors=["RED"], size=size)
         return pie_chart(
             title=title,
             data=(["PASSED", "FAILED"], [success_count, row_count - success_count]),

src/evidently/presets/dataset_stats.py

@@ -46,6 +46,7 @@
 from evidently.metrics.dataset_statistics import DuplicatedColumnsCount
 from evidently.metrics.dataset_statistics import EmptyColumnsCount
 from evidently.metrics.dataset_statistics import EmptyRowsCount
+from evidently.metrics.row_test_summary import RowTestSummary
 
 
 class ValueStats(ColumnMetricContainer):
@@ -468,10 +469,11 @@ def get_value_stats(self, context: Context) -> List[ValueStats]:
                 include_tests=self.include_tests,
             )
             for column in cols
+            if column not in (context.data_definition.test_descriptors or [])
         ]
 
     def generate_metrics(self, context: Context) -> Sequence[MetricOrContainer]:
-        metrics: List[MetricOrContainer] = [RowCount(tests=self._get_tests(self.row_count_tests))]
+        metrics: List[MetricOrContainer] = [RowTestSummary(), RowCount(tests=self._get_tests(self.row_count_tests))]
         value_stats = self.get_value_stats(context)
         metrics.extend(list(chain(*[vs.metrics(context)[1:] for vs in value_stats])))
         return metrics

tests/future/presets/test_test_fields.py

@@ -110,6 +110,8 @@ def value_stats_tests_check(metric: Metric, tests: Dict[str, ValueStatsTests]):
         return metric.tests == getattr(ts, field_name)
     if isinstance(metric, QuantileValue):
         return metric.tests == getattr(ts, f"q{int(metric.quantile * 100)}_tests")
+    if isinstance(metric, RowTestSummary):
+        return True
     raise ValueError(f"Unknown metric type {metric.__class__.__name__}")