import type * as PdfJs from 'pdfjs-dist'
import { pdfText, type PdfItem } from './docs.ts'

// A PDF's text, read in the view with pdf.js. The view's frame allows no worker, no eval, and no
// network, so pdf.js runs in this thread (its worker's handler set on globalThis, which it looks for
// before it would start a worker), with font loading off, from the bytes it is handed; pdf.js 6 has
// no eval. Both
// halves load on the first PDF: they are most of this plugin's size, and the plugin's definitions,
// which are this same code evaluated without a page, never reach them.

let loading: Promise<typeof PdfJs> | null = null

function pdfjs(): Promise<typeof PdfJs> {
  loading ??= (async () => {
    ;(globalThis as { pdfjsWorker?: unknown }).pdfjsWorker = await import('pdfjs-dist/build/pdf.worker.min.mjs')
    return await import('pdfjs-dist/build/pdf.min.mjs')
  })()
  return loading
}

export async function readPdf(bytes: Uint8Array, name: string): Promise<string> {
  const lib = await pdfjs()
  const task = lib.getDocument({ data: bytes, disableFontFace: true })
  try {
    const doc = await task.promise.catch((err: unknown) => {
      if ((err as { name?: string }).name === 'PasswordException') throw new Error(`${name} is password-protected.`)
      throw err
    })
    const pages: PdfItem[][] = []
    for (let n = 1; n <= doc.numPages; n++) {
      const page = await doc.getPage(n)
      pages.push((await page.getTextContent()).items as PdfItem[])
      page.cleanup()
    }
    const text = pdfText(pages)
    if (!text) throw new Error(`${name} has no text layer, as a scan has none; scans are not read yet.`)
    return text
  } finally {
    await task.destroy()
  }
}