Merge pull request #21142 from mvdbeek/new_datasets_when_creating_skipped_expression_tool_outputs

[25.0] Create new datasets when creating skipped database operation tool outputs

Author: mvdbeek

lib/galaxy/model/__init__.py

@@ -4948,9 +4948,15 @@ def get_quota_source_label(self):
 
     quota_source_label = property(get_quota_source_label)
 
-    def set_skipped(self, object_store_populator: "ObjectStorePopulator") -> None:
+    def set_skipped(self, object_store_populator: "ObjectStorePopulator", replace_dataset: bool) -> None:
         assert self.dataset
         object_store_populator.set_object_store_id(self)
+        if replace_dataset:
+            replacement = Dataset(state=Dataset.states.NEW)
+            replacement.object_store_id = self.dataset.object_store_id
+            self.dataset = replacement
+            self.dataset_id = None
+            self.dataset.object_store.create(self.dataset)
         self.extension = "expression.json"
         self.state = self.states.OK
         self.blurb = "skipped"

lib/galaxy/tools/actions/__init__.py

@@ -737,7 +737,7 @@ def handle_output(name, output, hidden=None):
                 hdca.collection.mark_as_populated()
             object_store_populator = ObjectStorePopulator(trans.app, trans.user)
             for data in out_data.values():
-                data.set_skipped(object_store_populator)
+                data.set_skipped(object_store_populator, replace_dataset=False)
         job.preferred_object_store_id = preferred_object_store_id
         self._record_inputs(trans, tool, job, incoming, inp_data, inp_dataset_collections)
         self._record_outputs(job, out_data, output_collections)

lib/galaxy/tools/actions/model_operations.py

@@ -178,7 +178,6 @@ def _produce_outputs(
             object_store_populator = ObjectStorePopulator(trans.app, trans.user)
             for hdca in output_collections.out_collection_instances.values():
                 hdca.visible = False
-                # Would we also need to replace the datasets with skipped datasets?
                 for data in hdca.dataset_instances:
-                    data.set_skipped(object_store_populator)
+                    data.set_skipped(object_store_populator, replace_dataset=True)
         trans.sa_session.add_all(out_data.values())

lib/galaxy_test/api/test_workflows.py

@@ -5639,6 +5639,54 @@ def test_run_with_default_file_in_step_inline(self):
             content = self.dataset_populator.get_history_dataset_content(history_id)
             assert "chr1" in content
 
+    def test_conditional_skip_on_database_operation_collection_output(self):
+        with self.dataset_populator.test_history() as history_id:
+            summary = self._run_workflow(
+                """
+class: GalaxyWorkflow
+inputs:
+  input_collection: collection
+steps:
+  filter:
+    tool_id: __FILTER_FAILED_DATASETS__
+    in:
+      input: input_collection
+    when: $(false)
+""",
+                test_data="""
+input_collection:
+  collection_type: list
+  elements:
+    - identifier: el1
+      content: "test content 1"
+""",
+                history_id=history_id,
+                wait=True,
+                assert_ok=True,
+            )
+            invocation = self.workflow_populator.get_invocation(summary.invocation_id, step_details=True)
+
+            input_hdca_details = self.dataset_populator.get_history_collection_details(
+                history_id, content_id=invocation["inputs"]["0"]["id"]
+            )
+            input_hda = input_hdca_details["elements"][0]["object"]
+            filter_content = self.dataset_populator.get_history_dataset_content(
+                history_id=history_id, content_id=input_hda["id"]
+            )
+            assert "test content 1" in filter_content, f"Expected 'test content 1' in input, got: {filter_content}"
+
+            # Get the filter step output
+            filter_step = [s for s in invocation["steps"] if s["workflow_step_label"] == "filter"][0]
+            filter_output_id = filter_step["output_collections"]["output"]["id"]
+            hdca = self.dataset_populator.get_history_collection_details(history_id, content_id=filter_output_id)
+            hda = hdca["elements"][0]["object"]
+
+            # Assert that the filter output dataset contains the string 'null'
+            filter_content = self.dataset_populator.get_history_dataset_content(
+                history_id=history_id, content_id=hda["id"]
+            )
+            assert "null" in filter_content, f"Expected 'null' in filter output, got: {filter_content}"
+
     def test_conditional_flat_crossproduct_subworkflow(self):
         parent = yaml.safe_load(
             """