Serve the requested schema_version's columns from public v2 importmetadata

[rtibbles]

Overview

Studio's v2/importmetadata checks that schema_version is between 5 and 6, then always serves the current kolibri_public columns and labels the response with the requested version. A request for schema_version=5 gets v6 columns marked as v5. Kolibri clients will start requesting the version they can read (learningequality/kolibri#15358), so Studio should serve the columns of the requested version, as Kolibri's endpoint does.

Complexity: Medium
Target branch: hotfixes

Context

• Kolibri selects columns per version with for_version: import_metadata.py#L125-L147
• Studio's column selection: import_metadata_view.py#L151-L176

The Change

For each table, v2/importmetadata should return exactly the columns of the requested schema version, and the latest version when none is requested.

Acceptance Criteria

General

• For each supported schema_version (5 and 6), every table's rows have exactly the columns of Kolibri's frozen column map for that version.
• Without schema_version, the response is unchanged from today.
• An out-of-range or unparseable schema_version returns 400 with the same JSON body as Kolibri, e.g. ["Schema version is too high, exports only suported for versions 5 to 6"], instead of plain text.

Testing

• Tests cover each supported version.

AI usage

I asked Claude Code to compare Kolibri's and Studio's v2 public content APIs. It found this gap and drafted the sections. I set the scope and reviewed each section.

[rtibblesbot]

Plan: v2/importmetadata serves exactly the requested schema version's columns and returns Kolibri's JSON 400 bodies.

• JSON 400 bodies for bad schema_version
• Serve the requested version's columns
• Self-review the produced code

Assumptions — the issue did not settle these; say so if one is wrong.

• No schema_version serves v5 ∪ v6 columns (today's response) per AC 2; if "latest version" was meant, file_size goes and v5-schema clients lose file sizes.
• 05 and +6 normalise to 5/6 and are served; if they should 400, validate the raw string against EXPORT_SCHEMA_VERSIONS.

---

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

How was this generated?

• Implemented code following the pre-approved plan using test-driven development (red-green-refactor)
• Ran the project's full test suite and linter before committing
• For UI changes: captured live screenshots to verify rendering
• For backend changes: verified with example API/CLI output