Grants:Project/Rapid/Strainu/Enhance and secure the future of Romanian projects' critical infrastructure/Report

Report accepted

This report for a Rapid Grant approved in FY 2021-22 has been reviewed and accepted by the Wikimedia Foundation.

• To read the approved grant submission describing the plan for this project, please visit Grants:Project/Rapid/Strainu/Enhance and secure the future of Romanian projects' critical infrastructure.
• You may still comment on this report on its discussion page, or visit the discussion page to read the discussion about this report.
• You are welcome to Email rapidgrants at wikimedia dot org at any time if you have questions or concerns about this report.

Goals

Did you meet your goals? Are you happy with how the project went?

Yes, all the activites of the project were done and I learned a lot about the specificities of the ORES scores.

Outcome

Please report on your original project targets. Please be sure to review and provide metrics required for Rapid Grants.

Target outcome	Achieved outcome	Explanation
Writing the revert bot according to the bot request and comments revieved	The code is written, deployed and functional on ro.wp	During the development phase I received a lot of feedback, which allowed me to tweak the robot on the community's needs. One outcome that was not planned in this part was to have the robot patrolled very likely good changes, which account for about 25% of all unpatrolled changes.
Creating a notification framework that allows following the bot status on-wiki. This will include a semaphore-like status page for all the bots using it, including the notifier itself	A hardcoded notification bot was created and deployed.	This task took significantly longer than expected due to the fact that I could not use the toolforge-jobs wrapper to get information about my running jobs. This was a miss during the documentation phase and forced me to take a few shortcuts which make the coode not as useful as I would have liked.
Update the diacritics converter to allow several actions, such as notifying instead of moving. This is more appropriate to the current status of the Wiki	Not done on purpose.	After discussing the issue a bit with some colleagues, I decided that there is a high chance that the notifications will be ignored, so I let the actions as they were.
Update the deletion proposal bot with several deletion approaches.	Done.	The new approach is to propose articled for deletion each day. The number depends on the number of articles in category and the number of days in the month.
Deployment to Wikimedia Cloud Services	Done.	All the bots considered as critical are now run from toolforge.
submit presentations to regional conferences happening in 2022, such as the CEE Meeting	Done	My proposed presentation to the CEE Meeting was accepted.

Learning

Projects do not always go according to plan. Sharing what you learned can help you and others plan similar projects in the future. Help the movement learn from your experience by answering the following questions:

• What worked well?

  I believe the main thing that should be mentioned is the relatively high participation of the community in the bot request discussion for the ORES reverter. Compared with the other bots, there were a lot of opinions expressed, mostly in a constructive manner. The participation continued during the developing of the bot, which is also unusual. The Wikimedia infrastructure also proved reliable, allowing fast iterations without too many issues. The final result is satisfactory, with the number of reverts roughly in line with our initial expectations (based on ballpark calculations) and with the bot highly configurable from the wiki itself, without code intervention.

  For the rest of the existing bots the work was quite simple, with the switch to Toolforge totally uneventful.

• What did not work so well?
  • The monitoring bot proved much harder to build than expected. A combination of insufficient research and limited Kubernetes knowledge made it impossible to reach the expected results.
  • The ORES scores seem quite biased against anonymous users. I have logged a task for the ORES team to look into the issue.
• What would you do differently next time?
  • Group the code development tasks and the deployment tasks of the same project together. This is how I ended up doing them, and it makes total sense from a SDLC point of view.
  • Log some hours for work on the framework (pwb for this project). There were some improvements that could have been made and would have simplified the robot code significantly.

Finances

Grant funds spent

Please describe how much grant money you spent for approved expenses, and tell us what you spent it on.

This was a code-writing grant, so I present below the time spent on each of the items from the original proposal:

Activity	Proposed Hours	Actual Hours	Comments
Planning	0 h	8 h	this was part of my volunteer time
Development and maintenance of the revert bot	48 h	48 h	This bot's development was pretty much seamless. The original targets took a bit less than planned, but I spent the rest developing the auto-patrolling feature and fine-tuning the configuration file.
Development and maintenance of the existing bots	12 h	16 h	The reporting bot took an insane amount of time (~8h) for results under the ones expected. Not working on the diacritics robot saved ~2h.
Setting up the cloud instance	4 h	4 h	The toolforge documentation are very detailed by Wikimedia standards and allowed me to set this up without any issues.
Documentation and dissemination	0	TBD	The CEE meeting presentation hasn't been done yet. This is volunteer time.
Reserve fund	4 h	all spent	Spent on the reporting bug.
Total	68 hours	68 paid hours + volunteer time

Remaining funds

Do you have any remaining grant funds? No.

Anything else

Anything else you want to share about your project?