Repository navigation
Older-Kolibri Perseus export of QTI text-entry items keeps only the first accepted answer #6221
Description
❌ This issue is not open for contribution. Visit Contributing guidelines to learn about the contributing process and how to find suitable issues.
Target branch: unstable
Observed behavior
perseus_derive._derive_text (contentcuration/utils/assessment/qti/perseus_derive.py) builds the Perseus answers from <qti-correct-response> only and ignores <qti-mapping>.
Since #6187 the QTI editor writes a single-cardinality declaration whose <qti-correct-response> holds the first accepted answer; every accepted answer is in the <qti-mapping>. Publishing such a Numeric or Text entry item derives a Perseus question with one answer.
Errors and logs
N/A. No error; the derived item silently has fewer answers.
Expected behavior
The derived Perseus item lists every accepted answer: the <qti-map-entry> keys when a mapping is present, else the <qti-correct-response> values (legacy items).
User-facing consequences
Learners on older Kolibri versions, which render the derived Perseus item, are marked wrong for every accepted answer after the first.
Steps to reproduce
- In the QTI editor, create a Numeric question with accepted answers 1, 2, 3, and publish the channel.
- Import it into a Kolibri version that renders the Perseus derivation.
- Answer 2. It is marked incorrect.
Context
- Studio
unstableafter QTI editor emits cardinality="multiple" and match_correct for multi-answer Numeric and Text entry questions #6187 - Related: QTI editor emits cardinality="multiple" and match_correct for multi-answer Numeric and Text entry questions #6187, [QTI] Tracking issue -- UX enhancements #6103
AI usage
I used Claude Code to review #6187's branch. Claude traced the publish path and drafted this issue; I reviewed it against perseus_derive.py and the branch diff.
Activity
rtibblesbot commentedon Sep 30, 2026
Plan:
_derive_textlists every accepted answer —<qti-map-entry>keys when a<qti-mapping>is present, else<qti-correct-response>values.Assumptions — the issue did not settle these; say so if one is wrong.
unstable(a5dc590c5); tests hand-build its XML from the QTI editor emits cardinality="multiple" and match_correct for multi-answer Numeric and Text entry questions #6187 issue body — if the shipped shape differs, the fixture changes.<qti-map-entry>key is accepted regardless ofmapped-value— if authors can write 0-scored entries, they would wrongly become correct answers and need a filter.case-sensitiveis dropped, as today for correct values — if older Kolibri's Perseus input needs it, case variants are accepted wrongly.<qti-mapping>with no entries falls back to correct-response — if it should mean "no answers", empty-mapping items derive nothing.1.0, not1) — if older Kolibri compares numeric answers as strings, keys and typed input may mismatch, as for correct values today.@rtibblesbot's comments are generated by an LLM, and should be evaluated accordingly
How was this generated?