Time
S
Nick
Message
00:36 axfelix joined #dataverse
02:08 xvx joined #dataverse
03:26 axfelix joined #dataverse
04:43 garnett joined #dataverse
07:42 xvx joined #dataverse
07:51 xvx joined #dataverse
08:07 jri joined #dataverse
11:43 bencomp joined #dataverse
12:23 bmckinney joined #dataverse
12:33 xvx joined #dataverse
14:23 pameyer joined #dataverse
14:36 bmckinney joined #dataverse
15:18 pdurbin
bmckinney: I just sent you (and others) an email about file categories in DVN 3.x using this study as an example and I hope it makes sense: http://arc.irss.unc.edu/dvn/dv/odvn/faces/study/StudyPage.xhtml?globalId=hdl:1902.29/D-17236
15:28 pameyer joined #dataverse
15:34 iamtimmo joined #dataverse
15:34 pdurbin
pameyer: hi. Did my email make sense?
15:36 pameyer
pdurbin: let me find out - I'm still clearing the backlog
15:36 bmckinney
pdurbin: thanks, that's very useful
15:37 pdurbin
bmckinney: sure. It's kind of weird how it used to work in DVN 3.x but I think it's good to be aware of it.
15:37 pameyer
looks like it makes sense
15:37 pdurbin
I'm sure we'll get questions about it as people upgrade.
15:40 pameyer
can categories have attributes?
15:41 pdurbin
pameyer: I'm not sure if we even call them categories anymore in Dataverse 4 (except in the code).
15:41 pameyer
so it sounds like that might not help
15:41 pdurbin
pameyer: what do you mean by attribute?
15:42 pameyer
if one of the problems is mapping categories to directory path in zip export, only map categories with an attribute that indicates it's part of the path
15:42 pameyer
you could probably fit ordering in there as well
15:42 pameyer
does that make any sense?
15:46 pdurbin
oh, I see what you mean. yes, that might be helpful in an upgrade from DVN 3 scenario but if we don't worry about that I think the current suggestion is to add a new column/field (a string) to indicate the fully qualified path to the file
15:47 pameyer
adding another field is probably the simplest way to handle this.
15:48 bmckinney
are we talking about the filemetadata object?
15:48 pdurbin
buh. probably. one sec
15:48 pdurbin
this is a bit old: https://apitest.dataverse.org/guides/developers/database/schemaspy/tables/filemetadata.html
15:49 pdurbin
but yeah, it has "label" on it... which is the filename. so yeah, that table, probably
15:52 bmckinney
interesting that the schema doesn't explicitly link to the datafile.filesystemname (where we are putting the dataset-relative path for our current experiment)
15:53 pdurbin
bmckinney: I build that schemaspy page on phoenix on every build but I need to make it public.
15:54 bmckinney
pdurbin: very cool by the way, I was just telling Pete that I was wishing for something like that yesterday!
15:54 pdurbin
let me see if I can expose that today. knock it off my list
15:54 pdurbin
it's nice to see where "develop" is heading, schema wise
15:58 pdurbin
bmckinney pameyer: ok, this is from last Thursday: http://phoenix.dataverse.org/schemaspy/latest/relationships.html
16:03 * pdurbin
leaves a comment at https://github.com/IQSS/dataverse/issues/2513#issuecomment-199881517
16:03 bmckinney
pdurbin: where can we purchase the poster version!
16:04 pdurbin
heh. gonna need a bigger boat
16:16 pdurbin
pameyer bmckinney: hey have either of you heard of http://dbmi.hms.harvard.edu ?
16:18 pameyer
pdurbin: only in the context of hearing about a talk that might've been interesting, but happened a while before I'd heard about it
16:18 pameyer
so not usefully
16:19 pdurbin
ok. just wondering. this guy is a friend of mine: http://dbmi.hms.harvard.edu/person/visiting-scientists/steve-strassmann
16:23 pameyer
cool
16:24 pameyer
it's interesting how "data provenance and identity" start to mean different things in molecular biology
16:25 pdurbin
yeah
16:27 pdurbin
certainly something we think about: http://datascience.iq.harvard.edu/blog/aiding-reproducible-research-adding-provenance-data-citations
16:28 garnett joined #dataverse
16:32 axfelix joined #dataverse
16:46 metamattj joined #dataverse
17:26 axfelix joined #dataverse
18:19 pdurbin
bmckinney: if you want someone to beta test your bibtex feature: https://github.com/IQSS/dataverse/pull/3020#issuecomment-199937161 ... dunno if pameyer wants to advertise your installation yet though.
18:24 bmckinney
pdurbin: probably ok for him to test it. would be a good check to see if our multi-author separator (";") is acceptable.
18:25 pdurbin
yeah. sounds like a domain expert
18:30 jri joined #dataverse
18:31 pdurbin
bmckinney: that reminds me, ready for me to talk about an idea for making it unit-testable?
18:39 bmckinney
pdurbin: sure
18:40 pdurbin
bmckinney: here's the starting point: https://github.com/IQSS/dataverse/blob/v4.3/src/test/java/edu/harvard/iq/dataverse/export/ddi/DdiExportUtilTest.java
18:43 pdurbin
It exercises code at https://github.com/IQSS/dataverse/blob/v4.3/src/main/java/edu/harvard/iq/dataverse/export/ddi/DdiExportUtil.java
18:43 pdurbin
The key is the "DTO" objects like DatasetDTO and DatasetVersionDTO.
18:45 axfelix joined #dataverse
18:47 pdurbin
Those DTO objects are testable without having postgres running because (unlikey entities) they don't touch the database.
19:10 bmckinney
cool
19:10 pdurbin
:)
19:23 cnk joined #dataverse
20:20 bencomp joined #dataverse
20:35 pdurbin
I left a note about this at https://github.com/IQSS/dataverse/pull/3020#discussion-diff-57063062
21:27 bencomp joined #dataverse
21:32 axfelix joined #dataverse
22:54 iamtimmo joined #dataverse
23:43 xvx_ joined #dataverse
