Inspect PDF Objects

Use the low-level reader to inspect PDF dictionaries and their indirect objects. This is useful when identifying page resources before building a PDF optimizer. It does not extract or recompress image bytes; that requires handling each resource's stream and filters yourself.

import { createMuhammaraWasm } from "@muhammara/wasm";

function dereference(reader, object) {
  var reference = object.toPDFIndirectObjectReference();
  return reference ? reader.parseNewObject(reference.getObjectID()) : object;
}

function inspectPageXObjects(muhammara, bytes, pageIndex) {
  var reader = muhammara.createReader(bytes);

  try {
    var page = reader.parsePageDictionary(pageIndex);
    var resources = reader.queryDictionaryObject(page, "Resources");
    var xObjects =
      resources &&
      reader.queryDictionaryObject(resources.toPDFDictionary(), "XObject");

    if (!xObjects) {
      return [];
    }

    return Object.keys(xObjects.toPDFDictionary().toJSObject()).map(
      function (name) {
        var entry = xObjects.toPDFDictionary().queryObject(name);
        var reference = entry.toPDFIndirectObjectReference();
        var object = dereference(reader, entry);
        var dictionary = object.toPDFStream().getDictionary();
        var subtype = reader.queryDictionaryObject(dictionary, "Subtype");

        return {
          name: name,
          objectId: reference ? reference.getObjectID() : undefined,
          subtype: subtype ? subtype.toPDFName().value : undefined,
        };
      },
    );
  } finally {
    reader.end();
  }
}

var muhammara = await createMuhammaraWasm();
var bytes = new Uint8Array(await pdfFile.arrayBuffer());
console.log(inspectPageXObjects(muhammara, bytes, 0));

Page indexes are zero-based. queryDictionaryObject resolves a dictionary entry when it is an indirect reference; entries returned by queryObject() do not, so resolve those through their object ID before inspecting them. Reader-owned objects become invalid after reader.end(), so convert the properties you need to plain JavaScript values first.

Not every XObject is an image: Subtype can be Image, Form, or another PDF-defined type. Inspect stream dictionaries before assuming an object can be recompressed. See Low-Level Writer, Reader, And Modifier for reader lifecycle and API details.

The browser example's Inspect PDF tab counts each page's image and form XObjects this way, together with its Info metadata, text operations, annotations, and bookmarks, and renders the result as a one-page report.