Library of Congress Translator

kalbers · December 4, 2014

It seems as though the translator no longer works for the Library of Congress's new interface. Search results (eg, http://www.loc.gov/search/?q=baseball&sp=2) just return a Snapshot option, not a list of the results for import. Individual books are recognized as documents (eg http://lccn.loc.gov/10035921). And images are recognized only as Snapshots as well (eg http://www.loc.gov/pictures/item/npc2007019114/). Finally, searches in specific collections, such as Prints and Photography, return nothing at all (eg http://www.loc.gov/pictures/search/?q=baseball).

I poked around in the forums for info on this, but didn't come across anything. I'm having the same results on a machine running Windows 8 and Linux Mint. Windows is running FF34 and Zotero 4.0.24.1.

aurimas · December 4, 2014

Correct, the new interface is not yet supported.

kalbers · December 4, 2014

Thanks.

I'd actually be interested in helping on this. I've played around with Scaffold before, but I'm not exactly sure what the best way to go about it is. They seem to still be using Voyager, would an update to the existing translator make the most sense, or would it need a wholly new one?

adamsmith · December 4, 2014

you'd likely want a new translator, much simpler than the Voyager one. Grab the MarcXML which is available straight from the page and just call the MarcXML import translator. The ILO translator could server as a template (though they do some modifications on import that likely won't be necessary here, so this will be even simpler).

aurimas · December 4, 2014

I was actually thinking that we could fetch the LCCN from page/URL and then use the LoC search translator to fetch metadata. Getting the MarcXML data is not as obvious from http://www.loc.gov/pictures/item/npc2007019114/ and search result pages. This would also potentially allow us to use fewer HTTP queries when requesting multiple sets of metadata. Still need to look more into this though.

aurimas · December 4, 2014

Though given the simple implementation of LoC search translator, there's no need to reuse that code (and extend it for LCCN or whatever those numbers are). We can query LoC's SRU servers directly from the web translator.

adamsmith · December 4, 2014

did you check whether the SRU works with npc2007019114 type IDs? These aren't standard LCCNs I believe

aurimas · December 4, 2014

I'm looking into it. Currently, I don't see it working with standard LCCNs either, so I'm guessing I'm doing something wrong.

adamsmith · December 4, 2014

it's not inconceivable that we need to handle pictures entirely separately, which, given the marc available, wouldn't be the end of the world, either.

zuphilip · December 4, 2014

http://lx2.loc.gov:210/lcdb?version=1.1&operation=searchRetrieve&query="10035921"&startRecord=1&maximumRecords=5&recordSchema=marcxml as a start for the SRU...

adamsmith · December 4, 2014

as I thought, doesn't work with the picture IDs. I think we need to do these separately.

aurimas · December 4, 2014

The query should be http://lx2.loc.gov:210/lcdb?version=1.1&operation=searchRetrieve&query=bath.standardIdentifier=10035921&startRecord=1&maximumRecords=5&recordSchema=marcxml but yes, it doesn't work with "non-standard" LCCNs. I contacted LoC about this. Will give them a couple days to respond.