IQSS logo

IRC log for #dvn, 2013-07-30

We've moved! Please join #dataverse instead. The new logs are at http://irclog.iq.harvard.edu/dataverse/today

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

All times shown according to UTC.

Time S Nick Message
05:24 pdurbin joined #dvn
15:05 peterbull joined #dvn
16:10 peterbull joined #dvn
16:27 pdurbin peterbull: nice, you're starring SWORD stuff: https://github.com/pjbull?tab=activity :)
16:30 peterbull oh! I should store the DVN stuff too :)
16:30 peterbull star*
16:30 pdurbin :)
16:31 pdurbin peterbull: I gave you push access to https://github.com/dvn/swordpoc
16:31 peterbull great, i'll have an initial commit of the python client this afternoon
16:33 pdurbin awesome. I'm sticking with curl for now but who knows :)
16:36 peterbull oh, we'll get you hooked soon enough
16:36 pdurbin :)
17:25 peterbull joined #dvn
18:33 jwhitney joined #dvn
18:53 pdurbin jwhitney: oh good, you're here
18:53 pdurbin I've been working on getMediaResourceRepresentation :)
18:53 jwhitney pdurbin: Oh, nice
18:53 pdurbin well
18:53 jwhitney pdurbin: how's it going?
18:53 pdurbin let me push some code and I'll explain the problem
18:54 jwhitney ok
19:02 pdurbin jwhitney: ok, so: exploring getMediaResourceRepresentation, refactor auth · 709eefa · IQSS/dvn - https://github.com/IQSS/dvn/commit/709eefa
19:03 pdurbin what this boils down to is that I'm now doing a GET on /dvn/api/data-deposit/swordv2​/edit-media/hdl:1902.1/12345 (for example)
19:03 pdurbin hoping that this is the right place to show you a list of files belonging to a study
19:04 pdurbin but it seems like instead that's the place to download a "representation" of the study as a zip file
19:04 pdurbin jwhitney: so I still need to figure out where the spec says I can give you the information you want
19:05 pdurbin peterbull: if you have any ideas, please jump in :)
19:05 jwhitney ok, sec
19:05 pdurbin jwhitney: peterbull is writing a sword client in python for us! :)
19:05 jwhitney nice!
19:06 pdurbin jwhitney: anyway, I haven't forgotten about your request. I'll figure out a way to get a list of files to you
19:07 pdurbin jwhitney: also, dvn-build won't deploy this until midnight but I changed to the zip method we talked about
19:07 pdurbin previously, a PUT to the edit-media URI would add a file
19:08 jwhitney pdurbin: Ok, that's great.
19:08 pdurbin now it replaces the files with whatever is sent as a zip file
19:09 pdurbin my team isn't thrilled with this "solution" ... so it's subject to change... we think we'd prefer to let you pick and choose which files to replace... but for now... this is a way for you to basically "sync" the contents of OJS with DVN... to replace the files by sending a new zip file
19:09 pdurbin jwhitney: does this make sense?
19:10 jwhitney pdurbin: yes, it does -- and my only concern with sending single files is being able to describe those files, if file-level metadata's needed
19:10 pdurbin right
19:11 pdurbin well, this is all related... you need a list of files via SWORD
19:11 jwhitney pdurbin: otherwise, I'm all for it, since it is simpler
19:12 pdurbin yeah, simpler to "sync" ... to send a zip with the latest desired contents across
19:13 pdurbin but to us it feels a bit too brute force... and studies have "versions" when files change... this is all stuff I need to think some more about
19:13 jwhitney understood.
19:14 pdurbin cool
19:15 jwhitney also, your comment re: GET on EM-IRIs as a place to get a representation of the study -- I agree
19:15 pdurbin jwhitney: you agree that it should be a zip file?
19:15 jwhitney it seems the deposit receipt can contain a 2nd EM-IRI though -- apologies if you've already been trhough this
19:16 jwhitney with a feed type attr. --  <link rel="edit-media" type="application/atom+xml;type=feed" href="http://swordapp.org/em-IRI/43/my_deposit.feed"/>
19:16 jwhitney to get a feed representation of individual elements
19:18 pdurbin yeah, I agree that's what the spec says. but the return type for getMediaResourceRepresentation is MediaResource ... which seems to be binary as opposed to XML
19:19 pdurbin so for now the "binary" I'm returning is this: ByteArrayInputStream("FIXME: replace with zip of all study files".getBytes())
19:21 pdurbin anyway, I'll keep hacking on it
19:22 pdurbin jwhitney: I should mentione that peterbull is testing against dvn-build (the public web server I mentioned) so he'd be a good resource to ping if I'm not around
19:22 jwhitney ok great
19:24 pdurbin I mentioned to him that I may well insert a version number in the URL ... /dvn/api/data-deposit/v1/swordv2/ <-- "v1"
19:24 pdurbin and maybe v2 or v3 will implement all of SWORDv2
19:25 pdurbin jwhitney: but the idea being that v1 would have enough functionality for your plugin
19:25 jwhitney pdurbin: right on
19:26 pdurbin jwhitney: so is a list of files still the biggest thing you need right now?
19:27 jwhitney pdurbin: other than time, sleep, & a term paper, yes, that's most important
19:27 pdurbin another thing I could work on is the ability to update study metadata such as title or author ... in this case I think a "replace" would be fine... no need to target individual fields
19:28 jwhitney pdurbin: & let's keep talking about how zip vs single-file updates. If we can incorp. file metadata in the atom entry in a way that's not too clumsy, I'm all for the latter.
19:29 pdurbin hmm
19:29 pdurbin right now the atom entry I'm using doesn't talk about files at all
19:29 pdurbin jwhitney: you think maybe it should?
19:29 jwhitney pdurbin: no, DC is not a great option for this
19:29 jwhitney pdurbin: hierarchies relationships, etc. -- no dice
19:30 pdurbin I agree about dublin core
19:31 pdurbin let me work more on the listing of files
19:31 jwhitney pdurbin: ok.
19:31 pdurbin jwhitney: gotta pick up my kids soon (summer camp). get some sleep! :)
19:31 jwhitney pdurbin: thanks! :)

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

We've moved! Please join #dataverse instead. The new logs are at http://irclog.iq.harvard.edu/dataverse/today