Follow-up to #113, which deliberately scoped the web tool out ("Web tool can stay CSV/TSV-only for now") while the CLI got .xlsx/.json/.parquet support (faircode/loaders_extra.py, closed via #127).
That gap is now stale enough to be worth closing: profiler.html / assets/profiler-engine.js still only parse CSV/TSV text client-side, and the README has to carry an explicit caveat ("Excel .xlsx, JSON, and Parquet aren't supported client-side yet - use the CLI below").
Request:
.json should be easy - native JSON.parse handles both records and split orientations without a new dependency..xlsx/.parquet need an actual parsing library in the browser (e.g. SheetJS for xlsx, a WASM parquet reader) - worth scoping separately since it adds real bundle weight, unlike the JSON case.
Would land well as two PRs: JSON support now, xlsx/parquet as a follow-up once a library choice is made.
Follow-up to #113, which deliberately scoped the web tool out ("Web tool can stay CSV/TSV-only for now") while the CLI got
.xlsx/.json/.parquetsupport (faircode/loaders_extra.py, closed via #127).That gap is now stale enough to be worth closing:
profiler.html/assets/profiler-engine.jsstill only parse CSV/TSV text client-side, and the README has to carry an explicit caveat ("Excel.xlsx, JSON, and Parquet aren't supported client-side yet - use the CLI below").Request:
.jsonshould be easy - nativeJSON.parsehandles bothrecordsandsplitorientations without a new dependency..xlsx/.parquetneed an actual parsing library in the browser (e.g. SheetJS for xlsx, a WASM parquet reader) - worth scoping separately since it adds real bundle weight, unlike the JSON case.Would land well as two PRs: JSON support now, xlsx/parquet as a follow-up once a library choice is made.