Fb0405

cdisc_rules_engine/dataset_builders/domain_list_with_define_builder.py

@@ -31,11 +31,11 @@ def build(self):
         records = []
         for define_item in all_define_metadata:
             domain_name = define_item.get("define_dataset_name", "")
+            filename = domain_files.get(domain_name)
             record = {
-                "domain": domain_name,
-                "filename": domain_files.get(domain_name),
                 **define_item,
+                "domain": domain_name if filename is not None else None,
+                "filename": filename,
             }
             records.append(record)
-
         return self.dataset_implementation.from_records(records)

cdisc_rules_engine/enums/domain_presence_values.py

@@ -4,3 +4,4 @@
 class DomainPresenceValues(BaseEnum):
     DATASET = "STUDY"
     RECORD = ""
+    DOMAIN = "N/A"

cdisc_rules_engine/enums/sensitivity.py

@@ -5,3 +5,4 @@ class Sensitivity(BaseEnum):
     DATASET = "Dataset"
     RECORD = "Record"
     GROUP = "Group"
+    STUDY = "Study"

cdisc_rules_engine/models/actions.py

@@ -65,6 +65,7 @@ def generate_dataset_error_objects(self, message: str, results: pd.Series):
         )
         if "domain presence" in self.rule.get("rule_type", "").lower():
             error_object.dataset = DomainPresenceValues.DATASET.value
+            error_object.domain = DomainPresenceValues.DOMAIN.value
             for error in error_object.errors:
                 error.dataset = DomainPresenceValues.DATASET.value
                 error.row = DomainPresenceValues.RECORD.value
@@ -177,6 +178,10 @@ def generate_targeted_error_object(  # noqa: C901
             errors_list = self._generate_errors_by_target_presence(
                 data, targets_not_in_dataset, all_targets_missing, errors_df
             )
+        elif self.rule.get("sensitivity") == Sensitivity.STUDY.value:
+            errors_list = self._generate_errors_by_target_presence(
+                data, targets_not_in_dataset, all_targets_missing, errors_df
+            )
         elif self.rule.get("sensitivity") == Sensitivity.GROUP.value:
             grouping_variables = self.rule.get("grouping_variables", [])
 

cdisc_rules_engine/rules_engine.py

@@ -1,6 +1,8 @@
 from copy import deepcopy
 from typing import Iterable, List, Union
 from dateutil.parser._parser import ParserError
+import traceback
+
 from business_rules import export_rule_data
 from business_rules.engine import run
 from cdisc_rules_engine.config import config as default_config
@@ -58,7 +60,7 @@
     ExternalDictionariesContainer,
 )
 from cdisc_rules_engine.models.sdtm_dataset_metadata import SDTMDatasetMetadata
-import traceback
+from cdisc_rules_engine.enums.sensitivity import Sensitivity
 
 
 class RulesEngine:
@@ -174,6 +176,8 @@ def validate_single_rule(self, rule: dict, datasets: Iterable[SDTMDatasetMetadat
                     )
                     if limit_reached:
                         break
+        if rule.get("sensitivity") == Sensitivity.STUDY.value:
+            results = self._collapse_to_study_result(results)
         return results
 
     def _update_total_errors_and_check_limit(
@@ -671,3 +675,15 @@ def handle_validation_exceptions(  # noqa
             message=message,
             status=ExecutionStatus.EXECUTION_ERROR.value,
         )
+
+    def _collapse_to_study_result(self, results: dict) -> dict:
+        """
+        For study sensitivity rules, collapse all per-dataset results into a single
+        study-level result using the first non-skipped result as the representative.
+        """
+        for key, dataset_results in results.items():
+            for result in dataset_results:
+                if result.get("executionStatus") != ExecutionStatus.SKIPPED.value:
+                    return {"study": [result]}
+        first_key = next(iter(results))
+        return {"study": results[first_key]}

resources/schema/rule/Rule_Type.md

@@ -192,7 +192,7 @@ all:
 
 One row per dataset defined in Define-XML:
 
-- `domain`
+- `domain` - The domain if the dataset exists, null otherwise
 - `filename` - The file name if dataset exists, null otherwise
 - `define_dataset_name`
 - `define_dataset_label`

resources/schema/rule/Sensitivity.json

@@ -10,6 +10,9 @@
     },
     {
       "const": "Group"
+    },
+    {
+      "const": "Study"
     }
   ],
   "markdownDescription": "Determines what level of granularity issues should be generated within the report"

resources/schema/rule/Sensitivity.md

@@ -1,3 +1,7 @@
+## Study
+
+Will report once per submitted data study. Lends itself to Domain Presence Check rule types as these are cross-study checks that do not involve within-dataset checks. Reports one result per failing row in the dataset generated by the rule type, collapsed to a single study-level result.
+
 ## Dataset
 
 Report one result per dataset generated by the `Rule Type`, where a dataset or record within the dataset matches the rule failure criteria

tests/unit/test_dataset_builders/test_domain_presence_define_builder.py

@@ -199,7 +199,7 @@
                         ],
                     },
                     {
-                        "domain": "SE",
+                        "domain": None,
                         "filename": None,
                         "define_dataset_name": "SE",
                         "define_dataset_label": "Subject Elements",
@@ -241,7 +241,7 @@
             pd.DataFrame(
                 [
                     {
-                        "domain": "AE",
+                        "domain": None,
                         "filename": None,
                         "define_dataset_name": "AE",
                         "define_dataset_label": "Adverse Events",
@@ -260,7 +260,7 @@
                         ],
                     },
                     {
-                        "domain": "DM",
+                        "domain": None,
                         "filename": None,
                         "define_dataset_name": "DM",
                         "define_dataset_label": "Demographics",
@@ -279,7 +279,7 @@
                         ],
                     },
                     {
-                        "domain": "SE",
+                        "domain": None,
                         "filename": None,
                         "define_dataset_name": "SE",
                         "define_dataset_label": "Subject Elements",
@@ -293,7 +293,7 @@
                         "define_dataset_variables": ["STUDYID", "USUBJID", "SESEQ"],
                     },
                     {
-                        "domain": "EC",
+                        "domain": None,
                         "filename": None,
                         "define_dataset_name": "EC",
                         "define_dataset_label": "Exposure as Collected",
@@ -344,7 +344,11 @@ def test_domain_list_with_define_dataset_builder(
     if expected_results.empty:
         assert result_df.empty, f"Expected empty DataFrame for {test_description}"
     else:
-        assert list(result_df.columns) == list(
+        assert set(result_df.columns) == set(
             expected_results.columns
         ), f"Columns do not match for {test_description}"
-        pd.testing.assert_frame_equal(result_df, expected_results, check_dtype=False)
+        pd.testing.assert_frame_equal(
+            result_df.reindex(columns=expected_results.columns),
+            expected_results,
+            check_dtype=False,
+        )

tests/unit/test_rules_engine.py

@@ -1050,7 +1050,7 @@ def test_rule_with_domain_prefix_replacement(mock_get_dataset: MagicMock):
                 {
                     "executionStatus": ExecutionStatus.ISSUE_REPORTED.value,
                     "dataset": "STUDY",
-                    "domain": "AE",
+                    "domain": "N/A",
                     "variables": ["AE"],
                     "message": "Domain AE exists",
                     "errors": [