From 991b49b78e72c8ec2ded74d8108f599bea71435e Mon Sep 17 00:00:00 2001
From: mvdbeek <m.vandenbeek@gmail.com>
Date: Thu, 23 May 2024 18:45:06 +0200
Subject: [PATCH] Check dataset state when attempting to acces dataset contents

---
 lib/galaxy/managers/datasets.py               | 19 +++++++++++++++++++
 .../webapps/galaxy/services/datasets.py       | 17 +++++++++++++++--
 lib/galaxy_test/api/test_datasets.py          | 11 +++++++++++
 3 files changed, 45 insertions(+), 2 deletions(-)

diff --git a/lib/galaxy/managers/datasets.py b/lib/galaxy/managers/datasets.py
index 1bbe86ee7021..18c9ae5fe0ed 100644
--- a/lib/galaxy/managers/datasets.py
+++ b/lib/galaxy/managers/datasets.py
@@ -488,6 +488,25 @@ def serialize_dataset_association_roles(self, trans, dataset_assoc):
             rval["modify_item_roles"] = modify_item_role_list
         return rval
 
+    def ensure_dataset_on_disk(self, trans, dataset):
+        # Not a guarantee data is really present, but excludes a lot of expected cases
+        if dataset.purged or dataset.dataset.purged:
+            raise exceptions.ItemDeletionException("The dataset you are attempting to view has been purged.")
+        elif dataset.deleted and not (trans.user_is_admin or self.is_owner(dataset, trans.get_user())):
+            raise exceptions.ItemDeletionException("The dataset you are attempting to view has been deleted.")
+        elif dataset.state == Dataset.states.UPLOAD:
+            raise exceptions.Conflict("Please wait until this dataset finishes uploading before attempting to view it.")
+        elif dataset.state == Dataset.states.DISCARDED:
+            raise exceptions.ItemDeletionException("The dataset you are attempting to view has been discarded.")
+        elif dataset.state == Dataset.states.DEFERRED:
+            raise exceptions.Conflict(
+                "The dataset you are attempting to view has deferred data. You can only use this dataset as input for jobs."
+            )
+        elif dataset.state == Dataset.states.PAUSED:
+            raise exceptions.Conflict(
+                "The dataset you are attempting to view is in paused state. One of the inputs for the job that creates this dataset has failed."
+            )
+
     def ensure_can_change_datatype(self, dataset: model.DatasetInstance, raiseException: bool = True) -> bool:
         if not dataset.datatype.is_datatype_change_allowed():
             if not raiseException:
diff --git a/lib/galaxy/webapps/galaxy/services/datasets.py b/lib/galaxy/webapps/galaxy/services/datasets.py
index fc67929c3c24..87d4da9c537a 100644
--- a/lib/galaxy/webapps/galaxy/services/datasets.py
+++ b/lib/galaxy/webapps/galaxy/services/datasets.py
@@ -364,7 +364,17 @@ def show(
         """
         Displays information about and/or content of a dataset.
         """
-        dataset = self.dataset_manager_by_type[hda_ldda].get_accessible(dataset_id, trans.user)
+        dataset_manager = self.dataset_manager_by_type[hda_ldda]
+        dataset = dataset_manager.get_accessible(dataset_id, trans.user)
+        requests_that_require_data = (
+            RequestDataType.converted_datasets_state,
+            RequestDataType.data,
+            RequestDataType.features,
+            RequestDataType.raw_data,
+            RequestDataType.track_config,
+        )
+        if data_type in requests_that_require_data:
+            dataset_manager.ensure_dataset_on_disk(trans, dataset)
 
         # Use data type to return particular type of data.
         rval: Any
@@ -606,7 +616,9 @@ def display(
         headers = {}
         rval: Any = ""
         try:
-            dataset_instance = self.dataset_manager_by_type[hda_ldda].get_accessible(dataset_id, trans.user)
+            dataset_manager = self.dataset_manager_by_type[hda_ldda]
+            dataset_instance = dataset_manager.get_accessible(dataset_id, trans.user)
+            dataset_manager.ensure_dataset_on_disk(trans, dataset_instance)
             if raw:
                 if filename and filename != "index":
                     object_store = trans.app.object_store
@@ -668,6 +680,7 @@ def get_metadata_file(
         TODO: Remove the `open_file` parameter when removing the associated legacy endpoint.
         """
         hda = self.hda_manager.get_accessible(history_content_id, trans.user)
+        self.hda_manager.ensure_dataset_on_disk(trans, hda)
         file_ext = hda.metadata.spec.get(metadata_file).get("file_ext", metadata_file)
         fname = "".join(c in util.FILENAME_VALID_CHARS and c or "_" for c in hda.name)[0:150]
         headers = {}
diff --git a/lib/galaxy_test/api/test_datasets.py b/lib/galaxy_test/api/test_datasets.py
index 3c8c9daf3420..9705d2d2b577 100644
--- a/lib/galaxy_test/api/test_datasets.py
+++ b/lib/galaxy_test/api/test_datasets.py
@@ -343,6 +343,17 @@ def test_display(self, history_id):
         self._assert_status_code_is(display_response, 200)
         assert display_response.text == contents
 
+    def test_display_error_handling(self, history_id):
+        hda1 = self.dataset_populator.create_deferred_hda(
+            history_id, "https://raw.githubusercontent.com/galaxyproject/galaxy/dev/test-data/1.bed"
+        )
+        display_response = self._get(f"histories/{history_id}/contents/{hda1['id']}/display", {"raw": "True"})
+        self._assert_status_code_is(display_response, 409)
+        assert (
+            display_response.json()["err_msg"]
+            == "The dataset you are attempting to view has deferred data. You can only use this dataset as input for jobs."
+        )
+
     def test_get_content_as_text(self, history_id):
         contents = textwrap.dedent(
             """\