Refractor to use DataCite REST API

[kjgarza]

The endpoint "https://search.datacite.org/api" is going to be deprecated in early 2019.

https://github.com/ropensci/rdatacite/blob/c525d6d28efb37f6c24595208815ec297643e0e5/R/dc_search.R#L9

rdatacite should be using any of the other api.datacite.org endpoints. https://support.datacite.org/v1.2/reference

[sckott]

thanks much @kjgarza

will do

[sckott]

@kjgarza what is the equivalent in the new setup to the old Solr service?

[sckott]

@kjgarza any thoughts?

[mfenner]

We will soon retire Solr, having replaced it with Elasticsearch. The way to interact with Elasticsearch is via the JSON REST API, which we have updated in January to support queries as well as DOI registrations. Documentation is at https://support.datacite.org/docs/api.

We would be happy to help refactoring rdatacite to work with the new API. One cool new functionality would be DOI registrations, but the users for this would be repository managers not researchers.

[sckott]

Thanks Martin.

One cool new functionality would be DOI registrations, but the users for this would be repository managers not researchers.

Would these people use R to do this? If so I can implement it, but if not, then maybe doesn't make sense here

[mfenner]

I would have liked to have that functionality about 6 months ago, when I registered about 60k DOIs using R. I decided to make it a two-step process with an intermediary JSON-LD file with all the metadata in schema.org format generated by R, and then a command lines script in ruby.

I guess this functionality can wait for later.

[sckott]

Okay, sounds good

[sckott]

putting off to next milestone.

@mfenner what is the base url for API v2? In the docs it just has https://api.test.datacite.org - maybe you could put what the base url is in the docs?

[katrinleinweber]

Omitting .test seems to work: https://api.datacite.org/dois/10.5438/0012

[mfenner]

The base URL is api.datacite.org, api.test.datacite.org is used for testing and you can‘t rely on specific DOIs being available. I suggest I give rOpenSci a Test account for proper testing of the package.

Best,

Martin

Katrin Leinweber notifications@github.com schrieb am Mo. 21. Okt. 2019 um 11:04:

Omitting .test seems to work: https://api.datacite.org/dois/10.5438/0012

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/ropensci/rdatacite/issues/24?email_source=notifications&email_token=AAAFU32REZVNUGEWN7CTJYTQPVO7VA5CNFSM4GM7YHLKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEBZNV2A#issuecomment-544398056, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAAFU3YHJOHM4MPFNNTZSULQPVO7VANCNFSM4GM7YHLA .

-- Martin Fenner DataCite Technical Director http://orcid.org/0000-0003-1419-2405

[sckott]

sounds good to me, thanks

[sckott]

to further clarify, i know Solr is going away, replaced with Elasticsearch. Is there still search capability? it looks like the /dois route has a query parameter. I don't see any docs for what that accepts. is it just simple search, or does it support Dismax like search?

[mfenner]

The query syntax with Elasticsearch is slightly different, but also very powerful, see https://support.datacite.org/docs/datacite-search-user-documentation and https://www.elastic.co/guide/en/elasticsearch/reference/current/query-dsl-query-string-query.html#query-string-syntax.

[sckott]

thanks, makes sense that query string is supported and not the ES body queries

[sckott]

@mfenner some errors, any thoughts?

• /activities
  • https://api.datacite.org/activities?query=ecology returns error 500 with no useful message
  • https://api.datacite.org/activities?ids= returns no data - and activities/<id> doesn't work either
• /events - whereas I think most routes in the meta slot have totalPages, this route has total-pages

[sckott]

@katrinleinweber re-worked the pkg, can install at remotes::install_github("ropensci/rdatacite@refactor") - OAI fxns are gone, and works/members/data centers fxns are gone. let me know any feedback soon - i have to submit a new version early next week

[katrinleinweber]

Thanks for adapting this package to the new API :-) I can't promise to test-run my analysis with the new version that quickly, because I'll have to migrate away from dc_works() first.

• /dois
  • api.datacite.org/dois/10.7966%2F9h05-8h60 returns a 404 error, no DOIs seem to work with this route

https://doi.org/10.7966%2F9h05-8h60 returns "DOI Prefix [10.7966] Not Found", so that seems to the reason for the 404.

Other DOIs work fine for me: https://api.datacite.org/dois/10.5281/zenodo.1308060, https://api.datacite.org/dois/10.13140/rg.2.2.18135.01449

[sckott]

thanks @katrinleinweber for the clarification on /dois

[sckott]

done in #29

[kjgarza]

• /events - whereas I think most routes in the meta slot have totalPages, this route has total-pages

To get camelCase response one needs to call the V2 of the API. Just need to change the header to accept: application/vnd.api+json; version=2

[sckott]

@kjgarza thanks, ill try that

[sckott]

@kjgarza that does work for /events, but just noticed that it doesn't seem to influence the /reports route

curl https://api.datacite.org/reports | jq '.meta | keys'
[
  "page",
  "total",
  "total-pages"
]

curl -H 'Accept: application/vnd.api+json; version=2' https://api.datacite.org/reports | jq '.meta | keys'
[
  "page",
  "total",
  "total-pages"
]

curl -H 'Accept: application/vnd.api+json' https://api.datacite.org/reports | jq '.meta | keys'
[
  "page",
  "total",
  "total-pages"
]