Time
S
Nick
Message
00:18 pdurbin
yeah, it does look odd
08:11 jri joined #dataverse
11:20 jri joined #dataverse
12:32 danmcp joined #dataverse
13:58 danmcp joined #dataverse
16:05 danmcp joined #dataverse
16:05 danmcp
pdurbin: ping
16:05 danmcp
pdurbin: sorry was disconnected, not sure if you responded yet
16:26 pameyer joined #dataverse
16:29 pdurbin
danmcp: mornin. What's up
16:30 danmcp
pdurbin: question for you related to the group building an open service broker to connect to dataverse
16:30 danmcp
pdurbin: as someone consuming data through the dataverse api, would I want to connect to the whole dataverse deployment, or a single dataverse.
16:31 danmcp
pdurbin: this corresponds to how many choices a user will get in the OpenShift UI of what to connect to
16:32 danmcp
pdurbin: my thought was that it should show an option for each dataverse
16:35 pdurbin
danmcp: sigh. The term "dataverse" is overloaded. I stubbed out a glossary ( https://docs.google.com/spreadsheets/d/1xje8FADNjCKNBNExXjKQXrmRgc-WQxEv7jKJrDOuuLA/edit?usp=sharing ) but I think you get that Dataverse is what we call the software and within an installation there can be a hierarchy of dataverses.
16:36 danmcp
pdurbin: yeah, I am asking about the hierarchy
16:36 danmcp
pdurbin: does it make sense that a user would want to connect to just one of those for pulling data?
16:37 danmcp
pdurbin: or multiple is fine as well. but is it reasonable for them to connect to each one they want to pull data from?
16:37 danmcp
pdurbin: with an explicit choice
16:37 pdurbin
Sure. If you're an astronomer, perhaps you're interested in the hierarchy of data under the Center for Astrophysics dataverse: https://dataverse.harvard.edu/dataverse/cfa
16:38 pdurbin
If you're a social scientist, perhaps you're interested in the hierarchy of data under this social science dataverse (MRA): https://dataverse.harvard.edu/dataverse/mra
16:39 danmcp
pdurbin: cool, I might ask you the question next time we are on the phone together to make sure we are totally on the same page. But I am going to assume we understand each other for now :)
16:39 pdurbin
danmcp: sounds good. Do you need any guidance on which APIs to use?
16:39 danmcp
pdurbin: we will need the list dataverses api
16:39 danmcp
pdurbin: a pointer would be great
16:41 pdurbin
I think you should start with the "contents" API . It looks like this in the API Guide: GET http://$SERVER/api/dataverses/$id/contents
16:41 pdurbin
http://guides.dataverse.org/en/4.8.5/api/native-api.html#dataverses
16:42 danmcp
pdurbin: thx
16:42 pdurbin
sure. like this: https://dataverse.harvard.edu/api/dataverses/cfa/contents
16:45 danmcp
pdurbin: thx!
16:46 pdurbin
It's weird to me that these datasets about Roman warships are under the Center for Astrophysics dataverse but what do I know: http://dx.doi.org/10.7910/DVN/23994 :)
16:47 danmcp
pdurbin: actually, where is the top level list of dataverses?
16:50 pdurbin
danmcp: it's a little weird but it depends on the "alias" of the root dataverse. By default, this the alias is "root" but even on the demo site the alias of the root dataverse is "harvard". You can tell what the alias is by clicking on "Advanced Search" on the home page and inspecting the URL , such as https://demo.dataverse.org/dataverse/harvard/search
16:51 pdurbin
Here's one with the out-of-the box alias of "root" for the root dataverse: http://phoenix.dataverse.org/api/dataverses/root/contents
16:51 danmcp
pdurbin: cool thx
16:53 pdurbin
danmcp: please be advised, however, that there is a known performance problem with this "contents" API endpoint ( https://github.com/IQSS/dataverse/issues/2122 ) so I guess I'd suggest the Search API if it's too slow.
16:55 danmcp
pdurbin: cool thx
16:56 pdurbin
danmcp: also, if the use case is pulling data, like you said, you may want to use a standard that Dataverse supports called OAI-PMH. We also call it "harvesting". Harvesting "sets" can be defined on the server for various subject of interest and the client can harvest just that data. This is metadata only that's "harvested", not the files themselves.
16:57 pdurbin
But once you have the database IDs for the files, you can download them using the "access" API .
16:57 danmcp
pdurbin: cool, that will be the next step
16:58 pdurbin
Ok, so you do want to download the files themselves. That's fine.
16:59 pdurbin
Feedback on these APIs is great because I don't really use them myself except for testing.
16:59 danmcp
pdurbin: cool, we'll let you know how it goes
17:00 pdurbin
thanks. I'm still a little confused about what you're building but there's no rush
17:00 danmcp
pdurbin: yeah let's chat about it next meeting
17:00 pameyer
I'm happy to see another user interested in APIs
17:00 pdurbin
sounds good
17:01 pdurbin
pameyer: please keep the feedback coming :)
17:01 pdurbin
I see OpenShift at https://www.openservicebrokerapi.org so maybe that's what he was talking about when he said "open service broker".
17:10 danmcp joined #dataverse
17:42 pameyer joined #dataverse
18:31 pdurbin
ah, more on what Roman warships have to do with astrophyics: http://space.mit.edu/home/guenther/roman.html :)
18:32 pameyer joined #dataverse
18:39 pameyer joined #dataverse
18:45 pameyer
without reading either of those links; back when maritime navigation involved astronomy, I could image astrophysics being something the navy would be interested in (and possibly interested in funding research / keeping it secret)
18:51 pdurbin
yeah
19:02 pameyer joined #dataverse
21:20 pameyer joined #dataverse
22:14 pameyer
pdurbin: question for you about sysadmin types and dependencies
