37 lines
1.1 KiB
TypeScript
37 lines
1.1 KiB
TypeScript
import { ArchiveReadError } from '../errors'
|
|
import {
|
|
ARCHIVE_STAMP,
|
|
hasPdfHeader,
|
|
locateAttachment,
|
|
MAX_PDF_BYTES,
|
|
} from './scan'
|
|
|
|
/**
|
|
* The raw JSON a Faunosteo PDF carries, read from the bytes alone (BLK-009):
|
|
* no PDF object graph is parsed, so no object stream is ever inflated. pdf-lib
|
|
* stays on the export side.
|
|
*/
|
|
|
|
export { ATTACHMENT_NAME_PREFIX } from './scan'
|
|
export const PDF_KEYWORD = `${ARCHIVE_STAMP}/1`
|
|
|
|
/** Size cap first, then the `%PDF-` header: anything else is `invalid_pdf`. */
|
|
function assertPdfBytes(bytes: Uint8Array): void {
|
|
if (bytes.byteLength > MAX_PDF_BYTES || !hasPdfHeader(bytes)) {
|
|
throw new ArchiveReadError('invalid_pdf')
|
|
}
|
|
}
|
|
|
|
/**
|
|
* The raw JSON carried by the PDF, whatever its shape (bare report of any
|
|
* version or archive envelope), before any validation: what the archive
|
|
* check tool shows next to the read result. Throws ArchiveReadError
|
|
* (`invalid_pdf`, `no_attachment`, `attachment_stripped`).
|
|
*/
|
|
export async function readAttachedDocument(
|
|
bytes: Uint8Array,
|
|
): Promise<unknown> {
|
|
assertPdfBytes(bytes)
|
|
return (await locateAttachment(bytes)).data
|
|
}
|