From d4705870283176e7d2fd90f69ae49128217ecb6e Mon Sep 17 00:00:00 2001
From: mvdbeek <m.vandenbeek@gmail.com>
Date: Mon, 23 Sep 2024 17:19:45 +0200
Subject: [PATCH] Fix data_column ref to nested collection

---
 lib/galaxy/tools/parameters/basic.py | 12 +++++++----
 lib/galaxy_test/api/test_tools.py    | 30 ++++++++++++++++++++++++++++
 lib/galaxy_test/base/populators.py   |  4 ++--
 3 files changed, 40 insertions(+), 6 deletions(-)

diff --git a/lib/galaxy/tools/parameters/basic.py b/lib/galaxy/tools/parameters/basic.py
index 279adde7a92e..9eef495cb153 100644
--- a/lib/galaxy/tools/parameters/basic.py
+++ b/lib/galaxy/tools/parameters/basic.py
@@ -1457,8 +1457,8 @@ def get_column_list(self, trans, other_values):
             # Use representative dataset if a dataset collection is parsed
             if isinstance(dataset, HistoryDatasetCollectionAssociation):
                 dataset = dataset.to_hda_representative()
-            if isinstance(dataset, DatasetCollectionElement) and dataset.hda:
-                dataset = dataset.hda
+            if isinstance(dataset, DatasetCollectionElement):
+                dataset = dataset.first_dataset_instance()
             if isinstance(dataset, HistoryDatasetAssociation) and self.ref_input and self.ref_input.formats:
                 direct_match, target_ext, converted_dataset = dataset.find_conversion_destination(
                     self.ref_input.formats
@@ -1553,9 +1553,13 @@ def is_file_empty(self, trans, other_values):
         for dataset in util.listify(other_values.get(self.data_ref)):
             # Use representative dataset if a dataset collection is parsed
             if isinstance(dataset, HistoryDatasetCollectionAssociation):
-                dataset = dataset.to_hda_representative()
+                if dataset.populated:
+                    dataset = dataset.to_hda_representative()
+                else:
+                    # That's fine, we'll check again on execution
+                    return True
             if isinstance(dataset, DatasetCollectionElement):
-                dataset = dataset.hda
+                dataset = dataset.first_dataset_instance()
             if isinstance(dataset, DatasetInstance):
                 return not dataset.has_data()
             if is_runtime_value(dataset):
diff --git a/lib/galaxy_test/api/test_tools.py b/lib/galaxy_test/api/test_tools.py
index 78a4b3e05c92..56fea13f8285 100644
--- a/lib/galaxy_test/api/test_tools.py
+++ b/lib/galaxy_test/api/test_tools.py
@@ -2481,6 +2481,36 @@ def test_implicit_reduce_with_mapping(self):
             )
             assert output_hdca["collection_type"] == "list"
 
+    @skip_without_tool("column_multi_param")
+    def test_multi_param_column_nested_list(self):
+        with self.dataset_populator.test_history() as history_id:
+            hdca = self.dataset_collection_populator.create_list_of_list_in_history(
+                history_id, ext="tabular", wait=True
+            ).json()
+            inputs = {
+                "input1": {"src": "hdca", "id": hdca["id"]},
+                # FIXME: integers don't work here
+                "col": "1",
+            }
+            response = self._run("column_multi_param", history_id, inputs, assert_ok=True)
+            self.dataset_populator.wait_for_job(job_id=response["jobs"][0]["id"], assert_ok=True)
+
+    @skip_without_tool("column_multi_param")
+    def test_multi_param_column_nested_list_fails_on_invalid_column(self):
+        with self.dataset_populator.test_history() as history_id:
+            hdca = self.dataset_collection_populator.create_list_of_list_in_history(
+                history_id, ext="tabular", wait=True
+            ).json()
+            inputs = {
+                "input1": {"src": "hdca", "id": hdca["id"]},
+                "col": "10",
+            }
+            try:
+                self._run("column_multi_param", history_id, inputs, assert_ok=True)
+            except AssertionError as e:
+                exception_raised = e
+            assert exception_raised, "Expected invalid column selection to fail job"
+
     @skip_without_tool("column_multi_param")
     def test_implicit_conversion_and_reduce(self):
         with self.dataset_populator.test_history() as history_id:
diff --git a/lib/galaxy_test/base/populators.py b/lib/galaxy_test/base/populators.py
index 4d1c46ca5309..014880ebd1e6 100644
--- a/lib/galaxy_test/base/populators.py
+++ b/lib/galaxy_test/base/populators.py
@@ -2898,7 +2898,7 @@ def __create_payload(self, history_id: str, *args, **kwds):
         else:
             return self.__create_payload_collection(history_id, *args, **kwds)
 
-    def __create_payload_fetch(self, history_id: str, collection_type, **kwds):
+    def __create_payload_fetch(self, history_id: str, collection_type, ext="txt", **kwds):
         contents = None
         if "contents" in kwds:
             contents = kwds["contents"]
@@ -2920,7 +2920,7 @@ def __create_payload_fetch(self, history_id: str, collection_type, **kwds):
                     elements.append(contents_level)
                     continue
 
-                element = {"src": "pasted", "ext": "txt"}
+                element = {"src": "pasted", "ext": ext}
                 # Else older style list of contents or element ID and contents,
                 # convert to fetch API.
                 if isinstance(contents_level, tuple):