Hacker News Reader

Comment by dredmorbius | Hacker News Reader

Question for anyone from Anna's Archive or elsewhere: are catalogue metadata available from national library collections such as the British Library or US Library of Congress?

(I've ... worked a bit with LoC classification and subject headings data, of which publicly-available data are only available in PDF or wordprocessing (MS Word or Wordperfect, if memory serves) formats. Which is ... somewhat unfortunate.)

View on HN

Check out z39.50, you can use it to search and pull data from most national and university library catalogs. But also be prepared for some code archaeology along the way, as the protocol is around 50 years old.

https://en.wikipedia.org/wiki/Z39.50 https://z-brary.com/ https://github.com/asl2/PyZ3950/blob/master/PyZ3950/zoom.py

Depending on what you're looking for, a _lot_ more is being published by the Library of Congress as Linked Data nowadays, including LoC classification and subject headings. Check out https://id.loc.gov/.

That seems to be just metadata about metadata.

Where do you go to actually find the Classifications or Subject Headings themselves, in whatever format?

https://id.loc.gov/download/ offers bulk downloads of many of the vocabularies, including the subject headings. In addition, individual records of various types can be search for.

Thanks.

That seems to have Subject Headings, but not the Classification, unless I'm blind (a fairly high likelihood).

I've spent quite some time poking around the LoC website over the years without finding anything more accessible than the PDFs of the Library of Congress Classification. Are you aware of that being generally available?

Near as I can tell, the best that's available is starting with the class(es) you want (e.g., https://id.loc.gov/authorities/classification/H.html for social sciences) and crawling recursively to grab whatever serialization of the RDF you want to consume. There's no SPARQL endpoint for id.loc.gov, alas, but it beats resorting to the PDFs.

Thanks.

Reply on news.ycombinator.com