Merge pull request #20880 from jmchilton/validate_columns

Validate sample sheet column definitions in workflow definitions on backend.

Author: mvdbeek

lib/galaxy/dependencies/pinned-requirements.txt

@@ -82,7 +82,7 @@ greenlet==3.2.4 ; (python_full_version < '3.14' and platform_machine == 'AMD64')
 grpcio==1.74.0
 grpcio-status==1.74.0
 gunicorn==23.0.0
-gxformat2==0.20.0
+gxformat2==0.21.0
 h11==0.16.0
 h5grove==2.3.0
 h5py==3.14.0

lib/galaxy/workflow/modules.py

@@ -41,6 +41,7 @@
 from galaxy.model.dataset_collections import matching
 from galaxy.model.dataset_collections.query import HistoryQuery
 from galaxy.model.dataset_collections.type_description import COLLECTION_TYPE_DESCRIPTION_FACTORY
+from galaxy.model.dataset_collections.types.sample_sheet_util import validate_column_definitions
 from galaxy.schema.invocation import (
     CancelReason,
     FailureReason,
@@ -1166,6 +1167,9 @@ def validate_state(self, state: dict[str, Any]) -> None:
                 fields=fields,
             )
             collection_type_description.validate()
+        column_definitions = state.get("column_definitions")
+        if column_definitions:
+            validate_column_definitions(column_definitions)
         return None
 
     def get_runtime_inputs(self, step, connections: Optional[Iterable[WorkflowStepConnection]] = None):

lib/galaxy_test/api/test_workflows.py

@@ -7836,6 +7836,72 @@ def test_invalid_collection_input_rejected(self):
         assert r.status_code == 400
         assert "Invalid collection type:" in r.json()["err_msg"]
 
+    @skip_without_tool("multi_data_optional")
+    def test_invalid_sample_sheet_definitions_rejected(self):
+        valid_collection_type = """
+class: GalaxyWorkflow
+inputs:
+  input:
+    type: collection
+    collection_type: sample_sheet
+    column_definitions:
+    - type: string
+      name: condition
+      default_value: treatment
+      optional: false
+      restrictions: ["treatment", "control"]
+steps:
+  multi_data_optional:
+    tool_id: multi_data_optional
+    in:
+      input1: input
+"""
+        r = self._post("workflows", files={"archive_file": io.StringIO(valid_collection_type)})
+        assert r.status_code == 200
+
+        invalid_collection_type = """
+class: GalaxyWorkflow
+inputs:
+  input:
+    type: collection
+    collection_type: sample_sheet
+    column_definitions:
+    - type: stringx
+      name: condition
+      default_value: treatment
+      optional: false
+      restrictions: ["treatment", "control"]
+steps:
+  multi_data_optional:
+    tool_id: multi_data_optional
+    in:
+      input1: input
+"""
+        r = self._post("workflows", files={"archive_file": io.StringIO(invalid_collection_type)})
+        assert r.status_code == 400
+
+        invalid_collection_type = """
+class: GalaxyWorkflow
+inputs:
+  input:
+    type: collection
+    collection_type: sample_sheet
+    column_definitions:
+    - type: string
+      name: condition
+      default_value: treatment
+      optional: false
+      restrictions: ["treatment", "control"]
+      extra_key: not_allowed
+steps:
+  multi_data_optional:
+    tool_id: multi_data_optional
+    in:
+      input1: input
+"""
+        r = self._post("workflows", files={"archive_file": io.StringIO(invalid_collection_type)})
+        assert r.status_code == 400
+
     @skip_without_tool("random_lines1")
     def test_run_replace_params_over_default_delayed(self):
         with self.dataset_populator.test_history() as history_id:

pyproject.toml

@@ -43,7 +43,7 @@ dependencies = [
     "future>=1.0.0",  # Python 3.12 support
     "gravity>=1.1.1",
     "gunicorn",
-    "gxformat2",
+    "gxformat2>=0.21.0",
     "h5grove>=1.2.1",
     "h5py>=3.12",  # Python 3.13 support
     "httpx",