IQSS logo

IRC log for #dataverse, 2020-10-01

Connect via chat.dataverse.org to discuss Dataverse (dataverse.org, an open source web application for sharing, citing, analyzing, and preserving research data) with users and developers.

| Channels | #dataverse index | Today | | Search | Google Search | Plain-Text | plain, newest first | summary

All times shown according to UTC.

Time S Nick Message
08:07 icarito[m] joined #dataverse
08:07 yoh joined #dataverse
09:10 icarito[m] joined #dataverse
09:10 yoh joined #dataverse
09:31 sindri-baldur joined #dataverse
10:04 pvranckx joined #dataverse
12:00 Virgile joined #dataverse
12:13 donsizemore joined #dataverse
13:44 pdurbin joined #dataverse
14:07 Virgile joined #dataverse
14:08 dataverse-user joined #dataverse
14:19 pameyer joined #dataverse
15:23 pameyer if I get to play around with time travel, one of the things on my list would be an issue/PR to make the native api metadatablocks indexable by field name
15:24 pdurbin so you could search on "metadataBlock:citation"?
15:26 pameyer so that you could access without searching
15:27 pdurbin the citation block for "title" shows as "citation" here, if that helps: http://localhost:8080/api/admin/datasetfield/title
15:27 pameyer right now, you need to do [ key for (key,value) in fields if key==what_you_want] kinds of things
15:28 pameyer aka - metadata['citation']['author']
15:28 pameyer vs for field in metadata['citation'] ... if typeName=='author' ...
15:28 pameyer bad pseudo code, but maybe enough
15:29 pdurbin oh, indexable in the array sense
15:29 pdurbin or object sense
15:29 pameyer yeah
15:29 pameyer and changing it now would break compatability
15:30 pdurbin Have you played with the newish targeted edits of certain fields? Changing just the title, for example.
15:30 pameyer not seriously - was glad to see them, but haven't hit a case to use them yet
15:32 pdurbin After lunch, let's discuss more. I'd love to hear about your use case.
15:33 pameyer sure
16:43 juancorr joined #dataverse
17:25 pdurbin Ok, here is my attempt to help someone reset the dataverseAdmin password: https://groups.google.com/g/dataverse-community/c/8o2ntdCWnI4/m/q0Nf76ugCQAJ
17:25 pdurbin poikilotherm: yes, I saw https://github.com/IQSS/dataverse/issues/7290
17:26 pdurbin pameyer: ready to hear about your use case. Is it safe you assume you want access to your custom metadata blocks?
17:49 pameyer pdurbin: the case where it occured to me was, but it's something that seems generalizable to other cases
17:50 pdurbin I was thinking that some of the other output formats are eaiser to index into. Schema.org JSON-LD, for example.
17:50 pameyer maybe - hadn't tried looking at those
17:51 pameyer general idea is an external tool doing a GET request for a dataset, checking a field to see if it needs to take other action (update dataset metadata, etc)
17:53 pdurbin sure, that makes sense
17:53 pameyer the natural way to do it feels like dataset[citation][author] , or dataset[custom_block][custom_field_name]
17:53 pdurbin yeah
17:53 pdurbin this is related: https://github.com/IQSS/dataverse/issues/2110
17:54 pameyer from the title that does look similar
17:55 pameyer hadn't occured to me to try the exported json-jd
17:56 pdurbin The downside is that it won't have your custom fields.
17:56 pdurbin because it's based on a standard
17:57 pameyer .... from my experience with implementing json-ld on dataset landing pages into a format google's validator was happy with, "standard" isn't the term I'd use
17:59 pdurbin heh
17:59 pdurbin a flexible standard
18:00 pameyer that it is
18:04 pameyer my grumbling about the native api json syntax is more rought edges than bugs - things work, and it doesn't block anything.  but it raises the barrier for folks
18:04 pameyer so something I'd put into a v2/future blue sky API design
18:05 pameyer along with not requiring a repeat of the type defination in the value assignment for metadata fields
18:05 pdurbin See Raman's example of a cleaned up version at https://github.com/IQSS/dataverse/issues/3068
18:07 pameyer very much along those lines
18:07 pdurbin :)
18:08 pameyer Of course, an issue that mentions FRDs is one that I can recognize as being old without looking at the timestamps
18:31 pameyer pdurbin: https://groups.google.com/g/dataverse-community/c/8o2ntdCWnI4/m/q0Nf76ugCQAJ looks like a pretty clear walkthrough to me
18:33 pdurbin Thanks. And I'm glad you remember the FDRs. :)
18:33 pdurbin er FRDs*
18:36 pameyer the time before dbrooke arrived....
18:36 pameyer and then things became free of BRDs/FRDs ;)
18:39 pdurbin :)
18:47 icarito[m] joined #dataverse
18:47 yoh joined #dataverse
20:11 pameyer pdurbin: if you've got a moment, could you sanity-check my understanding of a search api thing?
20:18 pdurbin pameyer: of course
20:20 pameyer it's looking to me like there's no way to search for a dataset where a metadata field isn't set (since the field not having a value means it doesn't sent in the solr doc)
20:21 pameyer does that sound right to you?
20:21 pameyer I'm probably going to try some workarounds with dummy/empty values; but would rather ask before missing something
20:23 pdurbin So, my first thought is to get a dump from Solr to get a lay of the land. They try it in Solr. Then try it in Dataverse.
20:24 pdurbin to get a dump from Solr I use scripts/search/query
20:25 pameyer yup - I checked solr, that was when I learned that if you don't set a value for a field it it's not in the search document
20:25 pdurbin For me, with sample data loaded up, only one of my datasets uses alternativeTitle. You're asking how to find the dataset that don't have alternativeTitle?
20:26 pameyer yeah
20:27 pdurbin this seems to work: -alternativeTitle:*
20:27 pameyer :facepalm:
20:28 pameyer thank you!
20:28 pdurbin sure, hope it works for you
20:29 pameyer works in solr for me
20:30 pameyer will probably take a few rounds of working typos out to get it working with the rest in the fq's , but should work
20:33 pdurbin yeah, it's a little fiddly
20:33 pdurbin But the takeaway should be that if it works in Solr it should work in Dataverse.
20:38 pameyer I'd guess it's less fiddly than I am typo-y
20:39 pdurbin :)
20:40 pdurbin I guess I'm just thinking about how queries against Solr can get pretty complicated but we really don't do those.
20:44 pameyer keeping things as simple as possible usually helps
20:44 pdurbin yeah
21:03 pdurbin left #dataverse
22:17 dataverse-user joined #dataverse

| Channels | #dataverse index | Today | | Search | Google Search | Plain-Text | plain, newest first | summary

Connect via chat.dataverse.org to discuss Dataverse (dataverse.org, an open source web application for sharing, citing, analyzing, and preserving research data) with users and developers.