[shaohuzhang1]

refactor: update string filtering to be case-insensitive for improved search functionality

[shaohuzhang1]

Review of the diff

1. __contains → __icontains (lines 245, 400, 533)

Good change. Case-insensitive search is usually what users expect for a name filter.

⚠️ Potential issue: In PostgreSQL, icontains uses ILIKE '%...%', which cannot use a normal btree index. If these tables can become large, consider:

• A pg_trgm GIN index on asker.username / name, or
• Using full-text search (SearchVector) if search is a first-class feature.

Also note: Django's contains already escapes %/_, so no injection risk — just performance.

2. Content-Disposition f-string → plain string (lines 329, 500, 630)

Correct cleanup. The f-string has no placeholders, so removing it is safe and slightly faster/branch-free.

Minor suggestion: since a filename is being hardcoded repeatedly, consider extracting a constant or helper, e.g.:

def _xlsx_response(workbook):
    response = HttpResponse(content_type="application/vnd.ms-excel")
    response["Content-Disposition"] = 'attachment; filename="data.xlsx"'
    workbook.save(response)
    return response

This removes triplicate code and ensures future header changes (e.g. adding RFC 5987 filename*=UTF-8''... for non-ASCII names) only need to happen once.

💡 The MIME type application/vnd.ms-excel is for legacy .xls. Since you're writing .xlsx via openpyxl, the correct type is:

"application/vnd.openxmlformats-officedocument.spreadsheetml.sheet"

3. Added is_resource_filter = False (line ~408)

This aligns the branch with the sibling method at line ~536, where the same workspace_manage branch already sets the flag to False.

✅ Behavioral fix: when the user is a workspace manager, the extra "resource filter" step was previously still applied. Now it's skipped consistently. Confirm that this is intentional — before, managers were likely getting results filtered by _apply_resource_filter(...) even though they had full access, which would silently hide records. If tests exist for manager-scoped listing, they should be updated.

Summary table

Change	Verdict	Notes
icontains ×3	✅	Add trigram/GIN index if tables grow
Drop f-string in header	✅	Factor out duplicate into helper
is_resource_filter = False	✅ (verify)	Fixes inconsistency with sibling method
MIME type	⚠️	Should be xlsx-specific MIME

No correctness regressions spotted; the main open questions are the intent of the is_resource_filter change (tests?) and the MIME-type mismatch.