Time
S
Nick
Message
00:10 jri joined #dataverse
04:11 jri joined #dataverse
07:48 juancorr joined #dataverse
08:05 jri joined #dataverse
09:54 MrK joined #dataverse
11:45 pdurbin joined #dataverse
11:56 dataverse-user joined #dataverse
13:06 juancorr26 joined #dataverse
13:36 pdurbin
juancorr26: awesome that you've updated the Spanish translation: https://groups.google.com/d/msg/dataverse-internationalization-wg/yUNsTZ1FZ8k/f5tyVhPcBgAJ
13:41 donsizemore joined #dataverse
14:18 MrK joined #dataverse
14:25 juancorr26
pdurbin: Thanks, we are migrating to Dataverse v4.11 and we have now the opportunity to update the translation.
14:42 juancorr26
We are now investigating two issues related with the translation:
14:46 juancorr26
There are some problems with the facets. You can see it in the French and Spanish translation in our test server http://oaimadrono.uned.es:8080 . Some facets are not translated and another have "two names". "Auteur Nom" should be "nom de l'auteur" or "Auteur" in French. Same in Spanish
14:51 juancorr26
There are error traces parsing translated dates. Some dates have the translated value. e.g.,
14:51 juancorr26
[2019-03-05T15:44:33.334+0100] [glassfish 4.1] [SEVERE] [] [edu.harvard.iq.dataverse.util.DateUtil] [tid: _ThreadID=30 _ThreadName=http-listener-1(4)] [timeMillis: 1551797073334] [levelValue: 1000] [[
14:51 juancorr26
[2019-03-05T15:44:33.336+0100] [glassfish 4.1] [SEVERE] [] [] [tid: _ThreadID=30 _ThreadName=Thread-9] [timeMillis: 1551797073336] [levelValue: 1000] [[ java.text.ParseException: Unparseable date: "15-oct-2018" at java.text.DateFormat.parse(DateFormat.java:366) at edu.harvard.iq.dataverse.util.DateUtil.formatDate(DateUtil.java:44) at edu.harvard.iq.dataverse.search.SolrSearchResult.getDateToDisplayOnCard(SolrSear
14:53 isullivan joined #dataverse
15:01 pdurbin
juancorr26: "Subject" seems to always be in English.
15:03 pdurbin
isullivan: I was very happy to see https://github.com/IQSS/dataverse-client-python/pull/51
15:05 juancorr26
Yes it is. "Dataverse Category" and "Publication Year" are always in English too.
15:05 pdurbin
Why?
15:07 juancorr26
I don't know yet. We have used the dataverse-language-packs develop branch
15:08 dataverse-user joined #dataverse
15:09 pdurbin
Ok. I don't know either. Is https://github.com/GlobalDataverseCommunityConsortium/dataverse-language-packs/issues the right place to create an issue?
15:10 pdurbin
dataverse-user: welcome
15:10 MrK joined #dataverse
15:11 dataverse-user left #dataverse
15:11 juancorr26
I don't know. I think that it could be a Dataverse issue.
15:12 fgb62 joined #dataverse
15:12 juancorr26
I think that this week we can found it.
15:15 pdurbin
Ok. Good luck finding it. Please reach out if you need help.
15:16 juancorr26
Perfect.
15:25 pameyer joined #dataverse
15:30 Jim__ joined #dataverse
15:32 Jim__
Good morning! @pdurbin - you mentioned to me the other day that you may have examples of google analytics files that go beyond the minimum to track page hits. Is that something you can point me too? I was hoping it would be possible to standardize how to capture hits on the new btn_download, etc. classes...
15:39 pdurbin
Jim__: mornin! I just commented on your open graph work, which seems awesome. Thanks for the pull request! I think I was mostly saying that it sounded like an opportunity to collaborate on GA config. Let me walk down the hall and ask people who know more about what we actually do in production for Harvard Dataverse.
15:50 pdurbin
Back. Are you still there?
15:54 Jim__
yep
15:57 pdurbin
I just got a nice brain dump from Mike about Google Analytics for Harvard Dataverse. Are you aware of "To track analytics beyond pageviews, style classes have been added for end user action buttons..." added in Dataverse 4.11? Please see http://guides.dataverse.org/en/4.11/admin/monitoring.html#monitoring-http-traffic-from-the-client-side
15:59 Jim__
Right - that's what I'm asking about. It doesn't seem like there's any code to actually have google use those buttons yet.
16:00 pdurbin
I might be misunderstanding you but btn-compute, btn-contact, etc, was added in https://github.com/IQSS/dataverse/commit/2edb0ab
16:01 Jim__
matomo seems to have an api to just list the css classes it should pick up. For google, I think one would have to attach listeners to those classes and fire events. I was going to try that but if someone has already done it, or thought more about it, or has similar code from before these buttons, then I didn't want to duplicate.
16:02 Jim__
Right -the buttons are there - does google automatically use them?
16:02 pdurbin
I don't know.
16:03 pdurbin
I have access to https://analytics.google.com for Harvard Dataverse but I have no idea what I'm doing in here. Is there a config file I can download and share with you?
16:03 Jim__
The original issue was from somewhere using matomo (its previous name), where I think, once the css classes are added, they can just directly/one line tell matomo to monitor those buttons.
16:04 Jim__
Yeah - the /var/www/dataverse/branding/analytics-code.html from the Web Analytics section.
16:04 Jim__
(has your code in it)
16:06 Jim__
Our is just the basic script that google suggests adding to your page in the analytics console, which afaik, doesn't do events/css classes, just page hits.
16:06 Jim__
I guess if you see any download or other events in your google console, that would suggest you have it set up. I only see page hits in ours.
16:08 pdurbin
I'm not sure where to look. I'm not sure what the console is.
16:09 pdurbin
Let me ask in standup. Starts in 5 minutes.
16:13 pdurbin
Jim__: meanwhile, sorry I missed the QDR coordination meeting (goodbye lunch for a colleague). Is there a google doc with notes?
16:14 Jim__
by console, I just meant the analytics.google.com site - I only see page hits there. When you're logged in, there's a behavior/events section that I think would be where button events would be. And under admin/property/tracking info/tracking code there's a few line script they suggest that just calls their gtag.js script.
16:15 Jim__
I think we mostly had updates in the doc from last time.
16:15 pdurbin
Ok. Is the doc public? Can I see it?
16:15 Jim__
(and it must have been someone else talking about having analytics examples...)
16:17 Jim__
Not sure if it's open or not but you've commented in it before, so you should have access.
16:40 pdurbin
I think I found it. I see some 2019 stuff in there.
16:41 pdurbin
Jim__: at standup I learned a little more about our GA setup.
16:41 pdurbin
It sound like we don't do much. I don't think we actually have much config to share.
16:41 pdurbin
And the config we set up is years old and may not even work.
16:42 pdurbin
Under "Conversions" then "Goals" we have a few (old) goals set up:
16:42 pdurbin
Goal 1: Clicking on sign up page
16:42 pdurbin
Goal 2: How many people create a Shib account by clicking \"Log-in\" first?
16:42 pdurbin
Goal 3: How many people create a Shib account by clicking \"Sign Up\" first?
16:42 pdurbin
... but I think that's it.
16:43 pdurbin
Jim__: so Danny, Mike and I agree that your best next step would be to post about what you're trying to achieve on the dataverse-community mailing list so more people can weigh in.
16:43 Jim__
OK - thanks. Hopefully not too much work then to get some basic logging off the new buttons at least - I'll see what I can do.
16:43 pdurbin
Sounds great. And we are definitely open to an issue and pull request if you figure something out.
16:45 pdurbin
I don't know if you and Venki were at the same community meeting but he's interested in the web analytics stuff: https://github.com/IQSS/dataverse/issues/4630
16:52 Jim__
not sure if I met him, but that's the issue I was looking at to understand where the new css classes came from and why...
16:55 pdurbin
ok but the recent work is discussed over in https://github.com/IQSS/dataverse/issues/4660
16:55 pameyer
if I'm remembering correctly, analytics might not have been a specific goal for the new css classes
16:56 pdurbin
that issue got a hooray emoji from https://github.com/nniiicc :)
16:58 pdurbin
pameyer: well, there were multiple goals. Mike has the scoop. We new way is more flexible, more free form. And hopefully those new event style classes in 4.11 will help Jim__ and Venki get farther with Google Analytics.
17:06 pameyer
pdurbin: thanks. a while back I had some discussions with Mike about long-term stuff for analytics
17:07 pdurbin
gotcha
17:07 pdurbin
pameyer: our Jenkins is broken, by the way. :( I assume yours is still working.
17:09 pdurbin
Jim__: Mike suggests looking at https://developers.google.com/analytics/devguides/collection/analyticsjs/events
17:12 pameyer
pdurbin: I never got configs/pipelines for dataverse in jenkins working
17:12 pdurbin
Oh. But you know how to install and configure Jenkins at least. :)
17:12 Jim__
Thanks for the pointers. I've done events before, so I have the basic idea. I think all that's needed is basically to grab the new buttons and add onclick events to send info to google (as in the discussion on social tracking).
17:12 pameyer
pdurbin: at least minimally
17:13 pdurbin
Jim__: ok, but for the social tracking, please be aware that https://github.com/IQSS/dataverse/issues/4538 is still open and Mike just linked me to it.
17:14 Jim__
- thanks
17:19 pdurbin
sure
17:21 pdurbin
Jim__: in those notes I saw Archivematica mentioned. This just in: Dataverse-Archivematica Integration Now Available for Testing: https://groups.google.com/d/msg/dataverse-community/liT6DmUK8hI/qmzUSwLyBAAJ
17:24 Jim__
yep - saw that. Thanks.
17:24 pdurbin
sure
17:25 pdurbin
Did everyone see "Increasingly slow feedback loop for developers, increasingly large WAR files"? https://groups.google.com/d/msg/dataverse-dev/UmsGzXsTkec/IPMIX_VdBwAJ
17:32 pameyer
yup
17:33 pdurbin
Good. Over the weekend I had a chat with a friend about it over at http://irclog.greptilian.com/friendlyjava/2019-03-03
18:37 jri joined #dataverse
19:09 donsizemore joined #dataverse
19:58 pameyer
pdurbin: https://github.com/IQSS/dataverse/issues/5602 makes me thing disk problems
19:58 pameyer
storage, inodes, space - not completely sure
19:58 pdurbin
oh?
20:00 pameyer
I vaguely think that jenkins does something equivalent to `cp -r` when it's starting a new job
20:01 pameyer
but the sysadmins for that box probably know better than me :)
20:01 pdurbin
Sure. We have a job from which we can run commands like `df` or whatever. Could try. Might not run.
20:03 pameyer
if it's not a blocker, it might be simpler to leave it to the folks with shell access
20:03 pdurbin
Yeah. I have other stuff I can work on.
20:32 dataverse-user joined #dataverse
20:57 pdurbin
pameyer: 47 GB free in /var which is where the Jenkins jobs seem to write. All of the filesystems listed in `df -h` seem to have plenty of space.
20:59 pameyer
huh - so it's not space then
21:01 pdurbin
and actually, I just tried to clone a Sphinx job like I did earlier and it worked this time
21:07 jri joined #dataverse
21:27 pdurbin
pameyer: they restarted Jenkins and I think everything is ok now: https://github.com/IQSS/dataverse/issues/5602#issuecomment-469858887 . Java had been updated.
21:32 pameyer
pdurbin: today I learned that jenkins reloads libraries from disk outside restarts
21:33 pdurbin
¯\_(ツ)_/¯
21:33 pdurbin
Jenkins is still very mysterious to me.
