@docture/loader-pdfjs

DocumentLoaderPdfJs is a docture DocumentLoader backed by pdfjs-dist (Mozilla PDF.js).

pnpm add @docture/loader-pdfjs

import { createExtractor } from "@docture/core";
import { DocumentLoaderPdfJs } from "@docture/loader-pdfjs";

const extractor = createExtractor({
  loaders: [new DocumentLoaderPdfJs()],
  strategies: [/* … */],
});

Capabilities

text: true, geometry: true, images: false. It reads, it does not render. For rendering, compose it with a rasterizer:

import { withRasterizer } from "@docture/core";
import { NapiCanvasRasterizer } from "@docture/raster-canvas";

withRasterizer(new DocumentLoaderPdfJs(), new NapiCanvasRasterizer({ dpi: 300 }));

Options

| Option | Default | | |---|---|---| | password | none | for an encrypted PDF | | maxPages | all | stop early; useful for a cheap classification pass | | lineTolerance | 0.5 | vertical band for grouping runs into lines, as a fraction of median run height |

Notes

• Coordinates are PDF points, top-left origin, y increasing downward. lines is sorted top-to-bottom, spans left-to-right.
• A scanned PDF yields form: "scanned" and empty text rather than nonsense, so a text-layer strategy is simply not eligible for it.
• pdf.js needs its legacy build in Node and a filesystem path (not a file:// URL) for its standard fonts. Both are handled here so you never see either failure. src/pdfjs.ts explains why.
• pdf.js is loaded with await import() at first use, so constructing the loader is free.