search

gbif / literature-ws

Apache License 2.0
0 stars 1 forks source link

DOI as a separately searchable field #3

Closed dnoesgaard closed 3 years ago

dnoesgaard commented 3 years ago

Please allow specific DOI searching e.g. /search?doi=10.1038/s41598-019-45491-7

mike-podolskiy90 commented 3 years ago

Is it possible that one DOI corresponds to several literature items as identifier? If not I suggest mapping to be GET /literature/{DOI}

dnoesgaard commented 3 years ago

Good question. One DOI should only correspond to one identifier. So perhaps you're right.

mike-podolskiy90 commented 3 years ago

I deployed that to dev dev is temporarily unavailable, deployed to UAT

dnoesgaard commented 3 years ago

There are some that don't work e.g.

10.1134/S1995425520050030 10.1093/sysbio/syaa080 10.17223/19988591/50/2 10.22541/au.160373183.39716449/v1 10.1111/2041-210X.13523

mike-podolskiy90 commented 3 years ago

Thank you for testing, I'll check those

mike-podolskiy90 commented 3 years ago

It seems that because of additional slashes and capital letters. Gonna commit an update to fix that

MattBlissett commented 3 years ago

I think you already know, but remember DOIs are case-insensitive.

mike-podolskiy90 commented 3 years ago

redeployed

dnoesgaard commented 3 years ago

Looking good. Doing some extensive testing now and will let you know the results shortly.

dnoesgaard commented 3 years ago

These are causing problems:

10.1603/0022-2585(2008)45[775:awrdfm]2.0.co;2 10.1579/0044-7447(2008)37[138:lbrapi]2.0.co;2 10.1603/0046-225x(2008)37[442:saoxgc]2.0.co;2 10.1639/0007-2745(2008)111[140:rlob]2.0.co;2 10.1898/1051-1733(2006)87[80:BICAOF]2.0.CO;2

(DOIs are horrible, I know, sorry...)

mike-podolskiy90 commented 3 years ago

Wow, those are indeed terrible. Never seen such DOIs. Thank you for discovering them, I'll fix those

dnoesgaard commented 3 years ago

Good news is that the remaining 9900+ DOIs I tested worked fine :)

mike-podolskiy90 commented 3 years ago

Since we have so many problems with this I decided to test all the DOIs. I discovered that we have several DOIs which are related to several literature items:

I also found 6 literature items with invalid DOI:

So I'm going to change this to GET /search?doi={doi} and we also have to do something with invalid ones

mike-podolskiy90 commented 3 years ago

@dnoesgaard seems fine now /literature/serach?doi={doi}

dnoesgaard commented 3 years ago

Thanks, Mikhail. This is very useful. I will make sure these are fixed.

dnoesgaard commented 3 years ago

(I edited your comment to keep track of the items)

dnoesgaard commented 3 years ago

Ug, that last one has 18 records. Yikes, that's bad...

mike-podolskiy90 commented 3 years ago

Can we close this?

dnoesgaard commented 3 years ago

Sure. I'll deal with the last one separately.