Restore QTI exercises when restoring a channel by rtibblesbot · Pull Request #6243 · learningequality/studio

rtibblesbot · 2026-09-30T22:39:25Z

Summary

restore_channel rebuilt questions only from an exercise's Perseus archive. QTI-only exercises restored with no questions; dual-published ones restored QTI item ids as assessment_ids, which broke publishing.

References

Fixes #6211

Reviewer guidance

When a node has a qti_zip file, questions restore from it and the exercise archive is ignored. Nodes with only an exercise archive restore as before.
Package items restore as type=qti with XML in raw_data; item images are attached as EXERCISE_IMAGE files.
A converted legacy item gets back the assessment_id its K… identifier encodes; a QTI editor item, whose id is in neither archive, gets a fresh one.
Items wrapping a Perseus custom interaction restore as perseus_question, keyed by the hex in perseus/<hex>.json.
Item order follows the manifest across both kinds (get_assessment_item_resources_from_manifest in imsmanifest.py).
Graphie svg + data json in packaged Perseus questions are rejoined into EXERCISE_GRAPHIE files.
An unreadable or unreachable package restores no items, with a warning; it does not fall back to the exercise archive, whose QTI ids cannot republish.

AI usage

Implemented by Claude Code following a pre-approved plan with TDD. Python tests ran in CI on fork verify branches (the agent's sandbox blocks Postgres): full suite passes on the branch and on its second commit alone.

@rtibblesbot's comments are generated by an LLM, and should be evaluated accordingly

How was this generated?

Ran pre-flight CI checks (lint, format, tests) and verified all pass
Rebased onto the target branch and resolved any conflicts
Reorganized commit history into clean, logical commits
Audited the diff to ensure only issue-relevant files are changed
Built PR body from the repository's PR template with evidence blocks

	🟡 Waiting for feedback Last updated: 2026-10-06 04:46 UTC

rtibblesbot · 2026-10-01T05:24:51Z

Outside this PR's scope — tracked under #5998:

[QTI] Restoring a published channel loses the questions of its units #6261 [QTI] Restoring a published channel loses the questions of its units

@rtibblesbot's comments are generated by an LLM, and should be evaluated accordingly

How was this generated?

Ran pre-flight CI checks (lint, format, tests) and verified all pass
Rebased onto the target branch and resolved any conflicts
Reorganized commit history into clean, logical commits
Audited the diff to ensure only issue-relevant files are changed
Built PR body from the repository's PR template with evidence blocks

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

rtibbles

More precise testing assertions would help here, but ultimately, not restoring in a broken way is key.

rtibbles

I am more confident that the code is actually doing what is intended in terms of setting the appropriate assessment_id - but the tests are too weak. If there are exceptions that need to be handled, that is fine, but we shouldn't weaken the assertions in the cases where the ids should be exactly replicated.

rtibbles · 2026-10-05T17:37:34Z

+                self.assertEqual(self._file_rows(item), self._file_rows(source_item))
+                if source_item.type == exercises.QTI:
+                    # An editor item's identifier is not its assessment_id.
+                    self.assertRegex(item.assessment_id, "^[0-9a-f]{32}$")


This assertion is too weak - it only asserts that the restored assessment_id is a valid 32 digit hex - not that it restored the original assessment id.

Fixed in bd281a4: an editor item whose identifier encodes its assessment_id (hex_to_qti_id) now must restore that exact id. Both node tests (QTI-only and dual-published) gained such an item.

Only an item named by an editor slug (e.g. item_1) keeps the regex check. Neither the package nor Kolibri's assessment_item_ids records its source assessment_id, so restore has to mint a new one.

Searched test_restore_channel.py for format-only assessment_id assertions and found 2:

_assert_restores_source_rows: changed as above.

test_editor_item_with_non_canonical_id_gets_a_new_assessment_id: left as is. Its identifier encodes no id, so there is no source id to compare against.

Exercises with a qti_zip restore every manifest item as type=qti with its images. A dual node's exercise archive holds QTI ids, so it is not a fallback: an unusable package restores no items. A converted legacy item gets back the assessment_id its identifier encodes. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

Packaged Perseus custom interactions restore with their hex assessment_id, images and graphies instead of landing as qti items. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

- Fixed fixture ids whose QTI encodings hold "-" and "_"; expected ids stated per test - Publish a restored legacy-only node through create_content_database Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

rtibblesbot · 2026-10-06T04:45:31Z

Outside this PR's scope — tracked in:

[QTI] Older Kolibri can't render exercises of only legacy items published from unstable #6294 [QTI] Older Kolibri can't render exercises of only legacy items published from unstable

Written by rtibblesbot, an LLM-based coding agent.

rtibbles

This now appears to properly encode and decode the assessment ids - this is good to go.

rtibblesbot · 2026-10-07T22:30:35Z

Quick check on rtibblesbot's work on this PR: react 👍 helpful · 😕 mixed · 👎 not helpful. Reply with anything specific.

rtibblesbot marked this pull request as ready for review September 30, 2026 23:56

rtibblesbot requested review from a team and rtibbles September 30, 2026 23:56

rtibblesbot force-pushed the issue-6211-939039 branch from 826690b to e919cbe Compare October 1, 2026 00:16

Read assessment item resources from QTI manifest

2287682

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

rtibblesbot force-pushed the issue-6211-939039 branch from 1324c5d to 2a49103 Compare October 1, 2026 06:13

rtibblesbot mentioned this pull request Oct 1, 2026

Leave answerless legacy input questions out of Perseus archives #6262

Open

rtibbles requested changes Oct 2, 2026

View reviewed changes

Comment thread contentcuration/contentcuration/utils/import_tools.py

Comment thread contentcuration/contentcuration/tests/test_restore_channel.py

rtibbles self-assigned this Oct 2, 2026

rtibblesbot force-pushed the issue-6211-939039 branch from 2a49103 to 5067b47 Compare October 2, 2026 20:49

rtibbles requested changes Oct 5, 2026

View reviewed changes

rtibblesbot force-pushed the issue-6211-939039 branch from 5067b47 to bd281a4 Compare October 6, 2026 03:22

rtibblesbot and others added 3 commits October 6, 2026 04:23

Restore Perseus questions packaged in QTI as perseus_question

65c85ad

Packaged Perseus custom interactions restore with their hex assessment_id, images and graphies instead of landing as qti items. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

Pin restored assessment ids and republished presets in restore tests

fa8e58f

- Fixed fixture ids whose QTI encodings hold "-" and "_"; expected ids stated per test - Publish a restored legacy-only node through create_content_database Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

rtibblesbot force-pushed the issue-6211-939039 branch from bd281a4 to fa8e58f Compare October 6, 2026 04:46

rtibbles approved these changes Oct 7, 2026

View reviewed changes

rtibbles merged commit 51657a6 into learningequality:unstable Oct 7, 2026
13 checks passed

rtibblesbot deleted the issue-6211-939039 branch October 7, 2026 22:30

This was referenced Oct 8, 2026

[QTI] Perseus archive of an exercise with raw Perseus questions keeps legacy input questions with no answer #6210

Open

Restoring a published QTI item stores lxml's reserialization instead of the published XML #6305

Open

Navigation Menu

Repository navigation

Restore QTI exercises when restoring a channel - #6243

Restore QTI exercises when restoring a channel#6243
rtibbles merged 4 commits into
learningequality:unstablefrom
rtibblesbot:issue-6211-939039

rtibblesbot commented Sep 30, 2026 •

edited

Loading

🟡 Waiting for feedback

Uh oh!

rtibblesbot commented Oct 1, 2026

Uh oh!

rtibbles left a comment

Uh oh!

Uh oh!

Uh oh!

rtibbles left a comment

Uh oh!

rtibbles Oct 5, 2026

Uh oh!

rtibblesbot Oct 6, 2026

Uh oh!

rtibblesbot commented Oct 6, 2026

Uh oh!

rtibbles left a comment

Uh oh!

Uh oh!

rtibblesbot commented Oct 7, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Conversation

rtibblesbot commented Sep 30, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

References

Reviewer guidance

AI usage

🟡 Waiting for feedback

Uh oh!

rtibblesbot commented Oct 1, 2026

Uh oh!

rtibbles left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

rtibbles left a comment

Choose a reason for hiding this comment

Uh oh!

rtibbles Oct 5, 2026

Choose a reason for hiding this comment

Uh oh!

rtibblesbot Oct 6, 2026

Choose a reason for hiding this comment

Uh oh!

rtibblesbot commented Oct 6, 2026

Uh oh!

rtibbles left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

rtibblesbot commented Oct 7, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

rtibblesbot commented Sep 30, 2026 •

edited

Loading