Merge pull request #273 from posit-dev/feat-step-report-col-vals-expr

fix: show a  `col_vals_expr()`-based step with the `get_step_report()` method

Author: rich-iannone

pointblank/_constants_translations.py

@@ -1843,6 +1843,33 @@
         "hi": "{column} Null नहीं है",
         "el": "{column} δεν είναι Null",
     },
+    "column_expr": {
+        "en": "The following column expression holds: {values}",
+        "fr": "L'expression de colonne suivante est vraie : {values}",
+        "de": "Der folgende Spaltenausdruck gilt: {values}",
+        "it": "La seguente espressione di colonna è valida: {values}",
+        "es": "La siguiente expresión de columna se cumple: {values}",
+        "pt": "A seguinte expressão de coluna é válida: {values}",
+        "ro": "Următoarea expresie de coloană este validă: {values}",
+        "tr": "Aşağıdaki sütun ifadesi geçerlidir: {values}",
+        "zh-Hans": "以下列表达式成立：{values}",
+        "zh-Hant": "以下欄位表達式成立：{values}",
+        "ja": "次の列式が成り立ちます：{values}",
+        "ko": "다음 열 표현식이 성립합니다: {values}",
+        "vi": "Biểu thức cột sau đây đúng: {values}",
+        "ru": "Следующее выражение для столбца выполняется: {values}",
+        "cs": "Následující výraz sloupce platí: {values}",
+        "pl": "Następujące wyrażenie kolumny jest spełnione: {values}",
+        "da": "Følgende kolonneudtryk gælder: {values}",
+        "sv": "Följande kolumnuttryck gäller: {values}",
+        "nb": "Følgende kolonneuttrykk gjelder: {values}",
+        "nl": "De volgende kolomuitdrukking geldt: {values}",
+        "fi": "Seuraava sarake-lauseke pätee: {values}",
+        "is": "Eftirfarandi dálktjáning gildir: {values}",
+        "ar": "التعبير التالي للعمود صحيح: {values}",
+        "hi": "निम्नलिखित स्तंभ अभिव्यक्ति सत्य है: {values}",
+        "el": "Η ακόλουθη έκφραση στήλης ισχύει: {values}",
+    },
     "rows_distinct_all": {
         "en": "All rows are distinct",
         "fr": "Toutes les lignes sont distinctes",
@@ -2032,6 +2059,33 @@
         "hi": "कॉलम <strong>{column_position}</strong> में {failure_rate} परीक्षण इकाई विफलताएँ",
         "el": "{failure_rate} ΑΠΟΤΥΧΙΕΣ ΜΟΝΑΔΩΝ ΔΟΚΙΜΗΣ ΣΤΗ ΣΤΗΛΗ <strong>{column_position}</strong>",
     },
+    "failure_rate_summary_no_column": {
+        "en": "{failure_rate} TEST UNIT FAILURES",
+        "fr": "{failure_rate} ÉCHECS D'UNITÉS DE TEST",
+        "de": "{failure_rate} TESTEINHEITENFEHLER",
+        "it": "{failure_rate} FALLIMENTI DI UNITÀ DI TEST",
+        "es": "{failure_rate} FALLOS DE UNIDAD DE PRUEBA",
+        "pt": "{failure_rate} FALHAS DE UNIDADE DE TESTE",
+        "ro": "{failure_rate} EȘECURI ALE UNITĂȚILOR DE TEST",
+        "tr": "{failure_rate} TEST BİRİMİ HATALARI",
+        "zh-Hans": "{failure_rate} 个测试单元失败",
+        "zh-Hant": "{failure_rate} 個測試單元失敗",
+        "ja": "{failure_rate} テストユニットの失敗",
+        "ko": "{failure_rate} 테스트 단위 실패",
+        "vi": "{failure_rate} LỖI ĐƠN VỊ KIỂM TRA",
+        "ru": "{failure_rate} СБОЕВ ТЕСТОВЫХ ЕДИНИЦ",
+        "cs": "{failure_rate} SELHÁNÍ TESTOVACÍCH JEDNOTEK",
+        "pl": "{failure_rate} NIEPOWODZEŃ JEDNOSTEK TESTOWYCH",
+        "da": "{failure_rate} TEST ENHED FEJL",
+        "sv": "{failure_rate} TESTENHETSFEL",
+        "nb": "{failure_rate} TESTENHETSFEIL",
+        "nl": "{failure_rate} TESTEENHEID FOUTEN",
+        "fi": "{failure_rate} TESTIYKSIKÖN VIRHEITÄ",
+        "is": "{failure_rate} PRÓFUNAREININGAR VILLUR",
+        "ar": "{failure_rate} فشل وحدات الاختبار",
+        "hi": "{failure_rate} परीक्षण इकाई विफलताएँ",
+        "el": "{failure_rate} ΑΠΟΤΥΧΙΕΣ ΜΟΝΑΔΩΝ ΔΟΚΙΜΗΣ",
+    },
     "not_shown": {
         "en": "(NOT SHOWN)",
         "fr": "(NON AFFICHÉ)",

pointblank/validate.py

@@ -15338,6 +15338,8 @@ def _step_report_row_based(
         text = STEP_REPORT_TEXT["column_is_null"][lang].format(column=column)
     elif assertion_type == "col_vals_not_null":
         text = STEP_REPORT_TEXT["column_is_not_null"][lang].format(column=column)
+    elif assertion_type == "col_vals_expr":
+        text = STEP_REPORT_TEXT["column_expr"][lang].format(values=values)
     elif assertion_type == "rows_complete":
         if column is None:
             text = STEP_REPORT_TEXT["rows_complete_all"][lang]
@@ -15384,10 +15386,14 @@ def _step_report_row_based(
         title = STEP_REPORT_TEXT["report_for_step_i"][lang].format(i=i) + " " + CHECK_MARK_SPAN
         assertion_header_text = STEP_REPORT_TEXT["assertion_header_text"][lang]
 
-        success_stmt = STEP_REPORT_TEXT["success_statement"][lang].format(
-            n=n,
-            column_position=column_position,
-        )
+        # Use success_statement_no_column for col_vals_expr since it doesn't target a specific column
+        if assertion_type == "col_vals_expr":
+            success_stmt = STEP_REPORT_TEXT["success_statement_no_column"][lang].format(n=n)
+        else:
+            success_stmt = STEP_REPORT_TEXT["success_statement"][lang].format(
+                n=n,
+                column_position=column_position,
+            )
         preview_stmt = STEP_REPORT_TEXT["preview_statement"][lang]
 
         details = (
@@ -15467,10 +15473,16 @@ def _step_report_row_based(
         assertion_header_text = STEP_REPORT_TEXT["assertion_header_text"][lang]
         failure_rate_metrics = f"<strong>{n_failed}</strong> / <strong>{n}</strong>"
 
-        failure_rate_stmt = STEP_REPORT_TEXT["failure_rate_summary"][lang].format(
-            failure_rate=failure_rate_metrics,
-            column_position=column_position,
-        )
+        # Use failure_rate_summary_no_column for col_vals_expr since it doesn't target a specific column
+        if assertion_type == "col_vals_expr":
+            failure_rate_stmt = STEP_REPORT_TEXT["failure_rate_summary_no_column"][lang].format(
+                failure_rate=failure_rate_metrics
+            )
+        else:
+            failure_rate_stmt = STEP_REPORT_TEXT["failure_rate_summary"][lang].format(
+                failure_rate=failure_rate_metrics,
+                column_position=column_position,
+            )
 
         if limit < extract_length:
             extract_length_resolved = limit

tests/test_validate.py

@@ -4669,6 +4669,53 @@ def test_col_vals_expr_pandas_tbl():
     )
 
 
+def test_col_vals_expr_step_report():
+    """Test that `get_step_report()` works for `col_vals_expr()` validations."""
+
+    # Polars test
+    df_pl = pl.DataFrame({"a": ["foo"], "b": ["bar"]})
+    validation_pl = Validate(data=df_pl).col_vals_expr(pl.col("a") == pl.col("b")).interrogate()
+
+    # This should not throw an exception (the original issue)
+    result_pl = validation_pl.get_step_report(1)
+    assert result_pl is not None
+
+    # Check that the expression is shown in the report
+    html_pl = result_pl.as_raw_html()
+    assert "The following column expression holds:" in html_pl
+
+    # Pandas test
+    df_pd = pd.DataFrame({"a": ["foo"], "b": ["bar"]})
+    validation_pd = Validate(data=df_pd).col_vals_expr(lambda df: df["a"] == df["b"]).interrogate()
+
+    # This should not throw an exception
+    result_pd = validation_pd.get_step_report(1)
+    assert result_pd is not None
+
+
+def test_col_vals_expr_display_text_formatting():
+    """Test that `col_vals_expr()` step reports don't show 'IN COLUMN None' text."""
+
+    # Create test data where expression will fail for some rows
+    df = pl.DataFrame(
+        {
+            "a": [1, 2, 3, 4, 5, 6],
+            "b": [2, 3, 4, 5, 6, 7],
+        }
+    )
+
+    # Test failing case, which should not show "IN COLUMN None"
+    validator_fail = Validate(df).col_vals_expr(pl.col("a") > pl.col("b")).interrogate()
+    report_html_fail = validator_fail.get_step_report(i=1).as_raw_html()
+
+    # Main assertions: no column references for expression validations
+    assert "IN COLUMN None" not in report_html_fail
+    assert "IN COLUMN" not in report_html_fail
+
+    # Verify it has the basic failure structure
+    assert "TEST UNIT FAILURES" in report_html_fail
+
+
 @pytest.mark.parametrize("tbl_fixture", TBL_LIST)
 def test_rows_distinct(request, tbl_fixture):
     tbl = request.getfixturevalue(tbl_fixture)