Handle DOI in CNKI and improve ID extraction

[zuphilip]

It will for one take the DOI from the website and save it as well, because it seems to be missing in the RefWorks data. Moreover, it is now possible to add a DOI as an identifer from CNKI, because I extended it to a search translator (similar to Airiti). It may slow down a little this functionality because CNKI takes its time to load, but I would say the advantages of supporting DOIs from CNKI are larger. Let me know what you think about this @dstillman @adam3smith .

(See also https://forums.zotero.org/discussion/80478/not-finding-digital-object-identifier-doi-for-chinese-texts)

[dstillman]

On https://forums.zotero.org/discussion/83483/possibility-of-getting-metadata-from-journals-published-in-china-with-dois @adam3smith mentions an API that we haven't been able to get to work. What's the deal with that?

In any case, I'm not sure about this, for both speed and privacy reasons — I don't think we can just send all DOIs people enter that don't otherwise resolve to a service described as being supported by the Propaganda Department of the Communist Party of China. It's CNKI's choice not to register these with doi.org, right? I think I've asked this before, but is there any other way to determine that a given DOI is from a given registrar?

It's possible we could somehow make this configurable, and off by default, though we've never done anything like that.

[dstillman]

is there any other way to determine that a given DOI is from a given registrar?

Answered here: https://forums.zotero.org/discussion/comment/357643/#Comment_357643

I don't know if this makes sense, but would it make sense to limit initial DOI lookup to DOI Content Negotiation and, instead of falling back to all DOI search providers, check and cache the registrar information from doi.org and only make direct requests based on that? (We'd need to change how search translators declared themselves as eligible for a DOIs, based on something from the given registrar info.)

[adam3smith]

I don't know if this makes sense, but would it make sense to limit initial DOI lookup to DOI Content Negotiation and, instead of falling back to all DOI search providers, check and cache the registrar information from doi.org and only make direct requests based on that?

I think that's right, yes. FWIW, doi.org will also return JSON via a RESTful API:
https://doi.org/api/handles/10.5064
That's likely faster&easier to parse than getting the html version.
(also didn't know this existed before Brenton posted the web version)

[dstillman]

doi.org will also return JSON via a RESTful API

Ah, nice.

I've created an issue for this.