I was writing an indexer (ca. 2018), and I don't recall encountering opaque blobs, but parsing the ZIP file and XML (with a small C XPath scanner) was straightforward.

But indexing PDFs, now there's a fun one.