New DatasetSet model
Merge request reports
Activity
changed milestone to %Arkindex 1.6.0
assigned to @mlbonhomme
added 4 commits
-
19a14db9...765221e3 - 3 commits from branch
master
- 879db112 - update models + data migration
-
19a14db9...765221e3 - 3 commits from branch
added 5 commits
-
20cc8b17...c09ad349 - 2 commits from branch
master
- 60064d02 - update models + data migration
- 86b7bf13 - wip API endpoints
- f0c38ccd - WIP fix API endpoints :'(
Toggle commit list-
20cc8b17...c09ad349 - 2 commits from branch
added 9 commits
-
f453ddcd - 1 commit from branch
master
- a1533a33 - update models + data migration
- 7912af40 - wip API endpoints
- 2864b10c - WIP fix API endpoints :'(
- b962b95e - fixed datasets API
- 52d8493d - temp dataset process fix
- d82903a4 - fix delete corpus
- 9662be71 - fix api schema
- e2c5b6e5 - put set back in destroydatasetelement
Toggle commit list-
f453ddcd - 1 commit from branch
requested review from @erouchet
This includes a temporary fix for Process Datasets because I really would like to put that in another merge request as this is already breaking my brain as is; similarly I'm going to create a new issue for the new dataset set mgmt API endpoints.
This looks like it works but there's still some very weird things that I can't figure out. For example, if you troubleshoot using the
arkindex.training.tests.test_datasets_api.TestDatasetsAPI.test_partial_update_ponos_task_state_update
test, you will notice that addingprefetch_related("sets")
here@extend_schema(tags=["datasets"]) class DatasetUpdate(ACLMixin, RetrieveUpdateDestroyAPIView): permission_classes = (IsVerified,) serializer_class = DatasetSerializer def get_queryset(self): queryset = Dataset.objects.filter(corpus__in=Corpus.objects.readable(self.request.user)) return queryset.prefetch_related("sets").select_related("corpus", "creator")
does not remove any queries but actually adds some, when there already are duplicated / triplicated / etc queries that retrieve dataset sets when using that endpoint (only when updating / when there is a Dataset instance, I think?) and I can't figure out why, at first I though maybe I had used something other than
dataset.sets.all()
somewhere but it actually looks like the prefetch_related doesn't work at all.mentioned in issue callico/callico#395 (closed)
added 13 commits
-
e2c5b6e5...4632b510 - 5 commits from branch
master
- 0229c525 - update models + data migration
- 5fe13227 - wip API endpoints
- 2a64e13d - WIP fix API endpoints :'(
- f2ced181 - fixed datasets API
- 5b2fcba9 - temp dataset process fix
- 94822ae1 - fix delete corpus
- b9b314c7 - fix api schema
- 99806a78 - put set back in destroydatasetelement
Toggle commit list-
e2c5b6e5...4632b510 - 5 commits from branch
added 1 commit
- e96d0a25 - regen fixtures + fix dataset process sets testing
added 12 commits
-
e96d0a25...f03910d6 - 3 commits from branch
master
- fdd611c8 - update models + data migration
- e64bc82d - wip API endpoints
- 3b4b7072 - WIP fix API endpoints :'(
- 7e4fc11d - fixed datasets API
- f5103b78 - temp dataset process fix
- b4f601b1 - fix delete corpus
- c87f3ed3 - fix api schema
- 58367d27 - put set back in destroydatasetelement
- 1b7da324 - regen fixtures + fix dataset process sets testing
Toggle commit list-
e96d0a25...f03910d6 - 3 commits from branch
added 1 commit
- c367fd09 - ensure temp process dataset sets fix migrations runs BEFORE the dataset set migration
mentioned in merge request !2264 (merged)
added 11 commits
-
c6405687 - 1 commit from branch
master
- 54cb800b - update models + data migration
- 2e89a321 - wip API endpoints
- ca9ea129 - WIP fix API endpoints :'(
- 39bd4dbd - fixed datasets API
- 0b9d79bf - temp dataset process fix
- 2217829b - fix delete corpus
- 9fc815d1 - fix api schema
- c35b0191 - put set back in destroydatasetelement
- ef55a9fc - regen fixtures + fix dataset process sets testing
- 760834d8 - ensure temp process dataset sets fix migrations runs BEFORE the dataset set migration
Toggle commit list-
c6405687 - 1 commit from branch
added 9 commits
- 0abeda52 - Make the set_elements field opt-in
- d67c18b0 - Add prefetch for both sets and set_elements
- f183c5c3 - Avoid filling up the RAM with dataset elements when cloning
- 9bf0e8d3 - Make the migration reversible
- 8727e94b - Add helpers to play with Django prefetches
- 6cafe501 - Restore set_elements on create responses
- f0543742 - Restore set_elements on update responses
- 9f41df7b - Update query counts
- 2df4601f - Always sort sets by name
Toggle commit listmentioned in issue #1698 (closed)