|
| 1 | +--- |
| 2 | +title: "Corpora builder gets its curator fixes — tags split on commas, and fetch stops clobbering your metadata" |
| 3 | +lede: "Two operator-reported curator bugs, fixed the same afternoon they surfaced: comma-separated tag entries now become separate chips instead of one fused mega-tag, and 'Fetch full content' no longer overwrites the title (and bib) you typed — it fills only empty fields and, when a page can't be read, says so instead of silently resetting the title to the URL. Plus the search-queue feature earns a splash highlight." |
| 4 | +date_created: 2026-08-02 |
| 5 | +date_modified: 2026-08-02 |
| 6 | +publish: true |
| 7 | +authors: |
| 8 | + - Michael Staton |
| 9 | +augmented_with: |
| 10 | + - Claude Code on Claude Opus 4.8 |
| 11 | +files_changed: |
| 12 | + - apps/strategy-curator/src/curation.svelte.ts |
| 13 | + - apps/strategy-curator/src/StrategyPicker.svelte |
| 14 | + - services/content-ingest/src/corpus.ts |
| 15 | + - splash/src/content/feature-highlights/fire-many-searches.md |
| 16 | +tags: |
| 17 | + - Augment-It |
| 18 | + - Strategy-Curator |
| 19 | + - Corpus |
| 20 | + - Bug-Fix |
| 21 | + - Tags |
| 22 | +--- |
| 23 | + |
| 24 | +# Corpora builder — curator fixes |
| 25 | + |
| 26 | +Two bugs surfaced live while curating the humain-vc theses, and both are fixed. |
| 27 | + |
| 28 | +## Tags split on commas (#76) |
| 29 | + |
| 30 | +Typing `Quantum-Computing, Quantum-Innovations, Computational-Biology` used to |
| 31 | +save as **one** dashed token, because the tag helper treated commas exactly like |
| 32 | +spaces. Now commas (and newlines) are tag **separators**; spaces within a |
| 33 | +segment stay word-joiners. Paste a comma-separated list into either tag field — |
| 34 | +source edit or corpus create — and you get distinct chips, deduped, each |
| 35 | +Train-Case. Suggestion clicks are unchanged. |
| 36 | + |
| 37 | +## Fetch full content stops clobbering your metadata (#77) |
| 38 | + |
| 39 | +"Fetch full content" used to overwrite the title you typed with the fetcher's |
| 40 | +guess — or, when the page couldn't be read, with the raw URL, which also hid |
| 41 | +that the fetch had failed. Enrichment is now **additive**: your saved title, |
| 42 | +publisher, date, and authors are authoritative; the fetch fills only fields you |
| 43 | +left empty and pulls the body. When a page can't be read, it now says |
| 44 | +"fetch failed — could not read the URL" instead of silently resetting. |
| 45 | + |
| 46 | +## Also included |
| 47 | + |
| 48 | +- The async **search queue** ("run several searches at once, no babysitting") |
| 49 | + earns a curated splash highlight, not just a changelog line. |
| 50 | + |
| 51 | +Both curator fixes reach augment.didi.sh on the next redeploy; already-mangled |
| 52 | +records (one quantum source's tags) were corrected directly the same day. |
0 commit comments