chenglou/pretext · error · Error

Unknown corpus ${id}. Available corpora: ${sources.map(sourc

Error message

Unknown corpus ${id}. Available corpora: ${sources.map(source => source.id).join(', ')}

What it means

Thrown by corpus-check.ts when --id was supplied but does not match any source.id in corpora/sources.json. Distinct from [30] (no id at all). The error re-lists all valid ids so a typo is immediately recoverable.

Source

Thrown at scripts/corpus-check.ts:303

  throw new Error(`Unsupported --method ${requestedMethod}; expected span or range`)
}
const overrideOptions: CorpusOverrideOptions = {
  font: parseStringFlag('font'),
  lineHeight: parseOptionalNumberFlag('lineHeight'),
  method: requestedMethod as 'span' | 'range' | null,
  sliceStart: parseOptionalNumberFlag('sliceStart'),
  sliceEnd: parseOptionalNumberFlag('sliceEnd'),
}
const sources = await loadSources()
const id = parseStringFlag('id')

if (id === null) {
  throw new Error(`Missing --id. Available corpora: ${sources.map(source => source.id).join(', ')}`)
}

const meta = sources.find(source => source.id === id)
if (meta === undefined) {
  throw new Error(`Unknown corpus ${id}. Available corpora: ${sources.map(source => source.id).join(', ')}`)
}

const lock = await acquireBrowserAutomationLock(browser)
const session = createBrowserSession(browser)
const diagnose = hasFlag('diagnose')

try {
  const port = await getAvailablePort(requestedPort === 0 ? null : requestedPort)
  const pageServer = await ensurePageServer(port, '/corpus', process.cwd())
  serverProcess = pageServer.process
  const baseUrl = `${pageServer.baseUrl}/corpus`
  console.log(`${meta.id} (${meta.language}) — ${meta.title}`)

  for (const width of getTargetWidths(meta)) {
    const requestId = `${Date.now()}-${width}-${Math.random().toString(36).slice(2)}`
    let url =
      `${baseUrl}?id=${encodeURIComponent(meta.id)}` +
      `&width=${width}` +

View on GitHub (pinned to ac49b09b7d)

Solutions

Copy the id verbatim from the error's availability list (or from corpora/sources.json).
Respect casing and hyphenation — the lookup is a strict equality on source.id.
If you added a new corpus, ensure it has an entry in corpora/sources.json before referencing it.

Example fix

// before
bun run scripts/corpus-check.ts --id=ja-kumo
// after
bun run scripts/corpus-check.ts --id=ja-kumo-no-ito

Defensive patterns

Strategy: validation

Validate before calling

const sources = await Bun.file('corpora/sources.json').json() as { id: string }[]
const VALID_IDS = new Set(sources.map(s => s.id))
function assertKnownCorpusId(id: string): void {
  if (!VALID_IDS.has(id)) {
    throw new Error(`Unknown corpus ${id}. Valid: ${[...VALID_IDS].join(', ')}`)
  }
}

Type guard

function isKnownCorpusId(id: string, known: Set<string>): boolean {
  return known.has(id)
}

Prevention

Copy the id verbatim from corpora/sources.json — the lookup is exact and case-sensitive.
When adding a corpus, register it in sources.json before referencing it from CI.
Avoid transcribing ids from prose/docs; they may be stale relative to the checked-in sources.json.

When it happens

Trigger: parseStringFlag('id') returns a value; sources.find(s => s.id === id) returns undefined. Triggers on a typo (`--id=ja-kumo`), a stale id from an older revision of sources.json, a case mismatch (`--id=JA-Kumo-No-Ito` — matching is exact and case-sensitive), or an id copied from FONT_MATRIX that is not (yet) in sources.json.

Common situations: Typos and case mismatches; using an id from a different branch where sources.json was edited; confusing a corpus id with its language tag or title.

Related errors

AI-assisted analysis of chenglou/pretext@ac49b09b7d (2026-08-12). Data as JSON: /api/errors/24fe0c7db07de0cc. Report an issue: GitHub.