Page MenuHomePhabricator

Citation logging splits events by citation type
Closed, ResolvedPublic

Description

In the visualeditorfeatureuse database we are getting separate window-open events for each citation type:

cite-any
cite-book
cite-krantenartikel
cite-libro
cite-Literatur
cite-nova
cite-online-news
cite-publicación periódica
cite-thesis
cite-Webseite
cite-מאמר

While this data isn't inherently useless, it isn't what we set out to record. And if at some point in the future we want to know what types of citations are being created, we should think harder about aggregation and localisation first.

For now it would be better if these were all tracked as 'cite'

Event Timeline

Esanders created this task.Feb 15 2019, 3:14 PM
Restricted Application added a subscriber: Aklapper. · View Herald TranscriptFeb 15 2019, 3:14 PM
Esanders updated the task description. (Show Details)Feb 15 2019, 3:15 PM

Change 490861 had a related patch set uploaded (by Esanders; owner: Esanders):
[mediawiki/extensions/Cite@master] Deprecated usage of generated cite-dialogs

https://gerrit.wikimedia.org/r/490861

Mvolz added a subscriber: Mvolz.Feb 15 2019, 5:39 PM

Elsewhere Neil said that this is useful for how the data is actually being analyzed currently, as he was just aggregating the various cite- values together.

Long-term, we could presumably switch away from the dialog-open logging into a dedicated feature:cite, action:[type] logging if we wanted to. As you say, though, localization is a concern there...

Change 490861 merged by jenkins-bot:
[mediawiki/extensions/Cite@master] Remove generated cite-dialogs stubs

https://gerrit.wikimedia.org/r/490861

Verified that we are now only seeing 'cite' in the feature column.

Verified that we are now only seeing 'cite' in the feature column.

Can you link to it?

ppelberg added a subscriber: ppelberg.EditedMar 13 2019, 9:25 PM

@Mvolz, the question you entered (below), does this come from a place of curiosity, from a place of wanting to verify the patch does what it's intended to do...something else entirely?

This ticket's presence in the "Product owner review" column leads me to assume this ticket is resolved and simply needing of an explicit declaration of such, but the comment below gave me pause, so here I am asking ¯\_(ツ)_/¯

Verified that we are now only seeing 'cite' in the feature column.

Can you link to it?

ppelberg added a subscriber: marcella.EditedMar 13 2019, 9:47 PM

"...if at some point in the future we want to know what types of citations are being created, we should think harder about aggregation and localisation first."

@Esanders + @DLynch, would you say the below is an accurate representation of the issue quoted in the line above?

Currently, it is difficult for us to aggregate window-open events for each citation type, because, as they are currently written, one citation type could be represented differently in different languages.

...just wanting to make sure I've understood you correctly. Also, please let me know if there is a better place to ask questions like these. Tagging @marcella
b/c this ticket came up during our Phabricator meeting today.

@Mvolz, the question you entered (below), does this come from a place of curiosity, from a place of wanting to verify the patch does what it's intended to do...something else entirely?

I've been thinking about doing more advanced stats for citoid the service to address T202001. I was thinking about introducing a user agent string to the citoid extension and compiling stats based on that, but maybe if this does the same thing that'd be superfluous?

It's also useful if we have usage statistics for the citoid page, see my reversion here: https://www.mediawiki.org/w/index.php?title=Citoid&type=revision&diff=3128634&oldid=3061360
Maybe that these stats can address that issue?

...I've been thinking about doing more advanced stats for citoid the service to address T202001. I was thinking about introducing a user agent string to the citoid extension and compiling stats based on that, but maybe if this does the same thing that'd be superfluous?

Ah, ok.

Considering the above [1] seems to be a comment/question about the the capabilities of this patch, rather than about whether the patch itself has been completed, I'm going to mark this task as "Resolved." If this is not how you understand the state of this patch, please let me know and I'll re-open it.


  1. "The above": I understood the comment/question above to resemble: Does this this patch enable us to compute summary measures about Citoid usage that do not include requests made by automated tools.
ppelberg closed this task as Resolved.Thu, Mar 28, 11:17 PM