On Fri, May 8, 2020 at 2:03 PM Shrinivasan T tshrinivasan@gmail.com wrote:

Hello all,

I am planning to build a grafana dashboard using Prometheus for the counts of all indic wiki articles.

The dashboards you are thinking of may actually already exist. Discovery of tools is a real problem in our community, and one I hope to be able to work on more in the coming months.

Please take a look at https://stats.wikimedia.org/#/ta.wikipedia.org/content/pages-to-date/normal|line|2-year|~total|monthly as an example of the data that the Wikimedia Foundation's Analytics team publishes to help folks keep track of trends across the Wikimedia movement's project wikis. More information on the "Wikistats 2" project can be found at https://wikitech.wikimedia.org/wiki/Analytics/Systems/Wikistats_2 including information on how you can contribute to this project.

Have to get all the counts and write a custom exporter.

Planning for a dashboard showing counts for articles in all indic languages.

Another dashboard to show counts for all wiki projects for selected language.

Have few queries.

1. How to get the number of articles in a wiki for example, tamil wikipedia

? Any api is there to get numbers?

Basic information on article counts can be fetched from each wiki using the Action API's action=query&meta=siteinfo endpoint. See https://www.mediawiki.org/wiki/API:Siteinfo for more information about this API.

See https://ta.wikipedia.org/wiki/%E0%AE%9A%E0%AE%BF%E0%AE%B1%E0%AE%AA%E0%AF%8D%E0%AE%AA%E0%AF%81:ApiSandbox#action=query&format=json&meta=siteinfo&siprop=statistics for an example usage on tawiki.

The Wikistats 2 project actually pulls its data from a public API as well! The dashboard I linked above fetches data from https://wikimedia.org/api/rest_v1/metrics/edited-pages/aggregate/ta.wikipedia.org/all-editor-types/all-page-types/all-activity-levels/monthly/2018033100/2020050800. This is part of what is known as the "Wikimedia REST API". See https://wikimedia.org/api/rest_v1/#/Edited%20pages%20data for more information on this API collection.

2. Can we run a sparql query from our own server?

3. Once these dashboards are built, can we host custom exporter, Prometheus

and grafana in tool server or any wiki cloud server? Whom to contact for hosting these ?

Toolforge is probably not a great place to host a Prometheus server simply because the local disk that you would have available to store the data sets would be hosted on the shared NFS server which provides $HOME directories for Toolforge maintainers and their tools.

A Cloud VPS project would be capable of hosting the general software described. See https://wikitech.wikimedia.org/wiki/Help:Cloud_VPS_project for more information about what a Cloud VPS project is and how you might apply to create one for your project.

Please be aware that a request to create the project described in this email would likely receive a response encouraging you to collaborate with the Wikistats 2 project to achieve your goals rather than making a new project.

Will do these in remote Hackathon this weekend.

I hope my answers here don't spoil your hackathon! Maybe try playing around with Wikistats 2 and the APIs it uses and think of ways that you could either add new features to Wikistats 2 or make a tool that uses data from the same APIs that would be helpful to the Indic language community.

Bryan