Page MenuHomePhabricator

For wiki pages and Wikidata items, make "Created' distinct from 'Improved'
Closed, ResolvedPublic2 Estimated Story Points

Description

For both Wikipedia pages and Wikidata items, the same item or page can currently be both as an Item Created and an Item Improved. (E.g., if a Page Created includes edits subsequent to initial creation, it is also counted as a Page Improved.)

Please separate these two categories so that they are distinct and mutually exclusive. I.e., for any individual wiki page or Wikidata item in an event, if Event Metrics counts it in the "Created" category, that page or Wikidata item can then never be also in the "Improved" category. The total of Created + Improved should equal the total of all pages involved.

Are the edits still counted? Yes! Edits made to Pages Created after the initial creation won't cause those pages to be numbered among the Pages Improved. But such edits are still counted in metrics like "Bytes changed" and "Edits" (in Event Summary) and "Edits during event" and "Bytes changed during event" (in Pages Created).

Metrics directly impacted by this ticket

Used in the Event Summary report and defined in T205561

  • Wikidata items created
  • Wikidata items improved
  • Pages created
  • Pages improved
  • Views to pages created
  • Avg. daily views to pages improved

Used in the Pages Improved report and defined in T210775

  • Title (which refers to the individual Pages Improved listed) This list should no longer include any articles also listed in the Pages Created report, T206058.

For testing

I created a couple of events that should be helpful:

Event Timeline

jmatazzoni updated the task description. (Show Details)
jmatazzoni added a subscriber: MaxSem.

@Mooeypoo please review this prior to Estimation. (I've put it on your list.) Thanks.

@jmatazzoni so, we want to not count edits that were done on items that were created during the event. Is that right?

If an item was created during an event, and then was edited several times -- we want to only count it as "item created" and ignore the improvements on it. Is that right? I just want to verify I understand correctly.

In T217455#4995118, @Mooeypoo wrote:

@jmatazzoni so, we want to not count edits that were done on items that were created during the event. Is that right?

If an item was created during an event, and then was edited several times -- we want to only count it as "item created" and ignore the improvements on it. Is that right? I just want to verify I understand correctly.

We want to keep Items Created and Items Improved distinct; A Wikidata Item Created cannot become an Item Improved, no matter how many times it is improved. I want to be clear, however, that we are not "ignoring" those improvements. Changes to Wikidata Items Created should be counted toward the totals in the Event Summary metrics Edits and Bytes changed.

it's the same with Pages Created and Pages Improved. A Page Created cannot become a Page Improved. But edits to Pages Created should be counted towards the totals in the Event Summary metrics above and in the Pages Created metrics below:

Edits during event
Bytes changed during event

jmatazzoni renamed this task from Make Wikidata "items created' distinct from 'items improved' to For wiki pages and Wikidata items, make "Items Created' distinct from 'Items Improved'.Mar 5 2019, 12:07 AM
jmatazzoni updated the task description. (Show Details)
jmatazzoni updated the task description. (Show Details)
jmatazzoni updated the task description. (Show Details)
MBinder_WMF set the point value for this task to 2.Mar 6 2019, 12:20 AM
MBinder_WMF added a subscriber: MBinder_WMF.

If this feels larger than the estimate, flag it to the team

From my own quick tests, this ticket is valid and the bug in fact does exist. It definitely worked before... yay regressions!

I went ahead and did this since I assume the bug would interfere with QA'ing other things.

PR: https://github.com/wikimedia/eventmetrics/pull/241

jmatazzoni renamed this task from For wiki pages and Wikidata items, make "Items Created' distinct from 'Items Improved' to For wiki pages and Wikidata items, make "Created' distinct from 'Improved'.Mar 22 2019, 3:16 PM

Merged. This should be easily QA'able, so moving to Product Sign-off, assuming that's okay!

I tested this using my two test events, which can both be found in the joe's testing events Program (I'm no longer providing links, since I've learned that URLs on Test are not stable). The two test events are:

  • 1 page created and then edited
  • 1 Wikidata item created then edited

In both cases, Event Metrics now shows 1 created, 0 improved. The number of Edits, however, is still correct, and includes both the original creation and the subsequent edits, as it should.

@dom_walden, I'm closing this ticket, but this is probably something we'll want to keep an eye on as we continue testing. Namely:

  • Does the system keep Pages/Items/Files Created and Improved distinct?
  • Does the system count edits properly for Pages Created?

The issue of whether we should edits to Wikidata items as "Edits" is one I will bring up elsewhere. Currently, we are not.