IQSS logo

IRC log for #dvn, 2013-11-13

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
01:06 axfelix joined #dvn
14:10 jwhitney joined #dvn
16:00 axfelix joined #dvn
16:41 JonathanNeal joined #dvn
17:47 pdurbin axfelix: hi. I hear you might have a mime type question
17:47 axfelix I sure might
17:48 axfelix nothing major -- elda raised the question of whether we should improve OJS notifications for when deposit to a linked DVN fails
17:48 axfelix and I wanted to check whether there's any reason deposit would fail other than HTTP issues
17:48 pdurbin fails? that's unpossible! ;)
17:48 axfelix e.g., failing a mimetype validation check on the DVN side
17:48 axfelix :) hey, we're tracking fails, we just wanted to see whether we should add more explicit error handling
17:49 pdurbin axfelix: does an error get spit out? (and what does it say?)
17:49 esotiri joined #dvn
17:50 axfelix I only have elda's testimony thus far but she claims she's having some deposits fail without any kind of error message from OJS, which is not expected behavior :)
17:50 axfelix I haven't asked her what files she's trying to provide
17:51 esotiri I have a R file that fails via ojs and directly on dvn (due to format..)
17:52 esotiri on dvn side we get an email for failure or, but nothing on ojs side
17:53 axfelix right, that's what I thought was going on -- we haven't anticipated failing due to mimetype since the existing OJS suppfile functionality which is effectively being replaced for journals which elect to use the DVN plugin allows anything other than potentially malicious files (afaik)
17:54 pdurbin well, OJS could "poll" the study by looking at the SWORD statement... "Contains feed of file entries, latestVersionState, locked boolean" -- http://thedata.harvard.edu/guides/dataverse-api-main.html#display-a-study-statement
17:54 axfelix in other words, I'm fairly certain there's a check in OJS preventing you from uploading php as a suppfile (because I remember having an emergency meeting about that several years ago), but something like R is obviously a grey area
17:56 axfelix is DVN's file format whitelist available anywhere?
17:57 pdurbin DVN accepts any file as "Other"
17:57 axfelix hm, then I wonder what the issue is...
17:57 axfelix jwhitney: you following this?
17:59 pdurbin axfelix: does "ingest" mean anything to you? it's what we do with "subsettable" files such as Stata and Rdata files
18:00 axfelix yup
18:00 axfelix kevin just clarified via email: "We accept all file types but certain files undergo additional processing and those are asynchronous and can fail based on the processing."
18:01 pdurbin so sword doesn't know what kind of file is being uploaded. it could be a tiny non-subsettable jpg or it could be a huge Rdata file
18:01 axfelix so that clarifies things to my satisfaction as far as what we need to do
18:02 pdurbin right, "ingest" is that extra processing... so we can visualize the data and do statistical analysis of it... only for "subsettable" files
18:02 pdurbin axfelix: if you query for the sword statement while ingest is going on, the study will indicate that it's locked
18:04 Kevin_QA joined #dvn
18:04 axfelix skype meeting for another project now, back in a bit
18:05 posixeleni joined #dvn
18:14 jwhitney Right now, on depositing files, the plugin examines the sword statement to verify files have been deposited but error handling can be improved.
18:16 pdurbin jwhitney: so you could detect only 2 files on the DVN side if you expect (after ingest) there to be 3
18:16 jwhitney pdurbin: yes. error handling for later deposits is better (adding a single file at a time)
18:17 jwhitney pdurbin: but on creation, right now only notifies  deposit success / fail, not individual files. This should be fixed.
18:17 pdurbin right. because the ingest might fail
18:18 pdurbin I guess I would just poll until I saw that the study is no longer locked
18:18 jwhitney Yes. Warns if study creation fails; warns if file(s) deposit fails, does not warn if something goes wrong w/ individual files
18:21 pdurbin ok. yeah. would be good to add that
18:22 pdurbin jwhitney: I don't know of a better way to tell you of a failure apart from the SWORD statement
18:23 pdurbin as esotiri mentioned, users get an email about how the ingest went... I guess in the OJS use case these emails would go to the journal manager
18:26 jwhitney pdurbin: yes, can use OJS's notificaiton system to warn submitter, etc. that file deposit has failed
18:26 pdurbin sounds great

| 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