Pixel Seed Data

This repository contains database backups that are to be used as seed data for Pixel. Since this repo is hosted on github pages, its files can be downloaded via https://wikimedia.github.io/pixel-seed-data/name-of-file.

Instructions

Let's say you're trying to add a new wiki page to Pixel. You would follow these steps:

1. Make sure you are on on the latest version of Pixel (including the latest database changes) and have a fresh copy of the database seed data by running ./pixel.js update (from Pixel's root directory and after you git checkout the main branch) before making any desired changes.
2. Run ./pixel.js reference to pull the latest code changes from MediaWiki. Then visit localhost:3000/wiki/<your-new-page> in your browser, and click the "Create" tab to create your page, etc. to do any other actions you wish that will affect the database.
3. In Pixel's root directory, run npm run db:save. This will generate a physical backup of the database Pixel uses and will save it in a folder called backups located in Pixel's root directory.
4. Drag and drop the tar file located in the backups directory anywhere on the page at https://github.com/wikimedia/pixel-seed-data. Github will upload the file and ask you to commit it to the main branch of this repo. Click "Commit changes".
5. IMPORTANT: Wait for the CI Job to pass before proceeding to the next step. It can take awhile (5 - 10 minutes) before the file is available to be downloaded.
6. Edit Pixel's Dockerfile.database file with the name of the tar file. Commit this change and create a pull request in Pixel to merge this change into Pixel's main branch.
7. After the pull request has been merged, locally checkout the main branch in Pixel and run ./pixel.js update to rebuild the Docker database image/volume with your changes.
8. Confirm your database changes are live by running ./pixel.js reference and visiting the server Pixel sets up at localhost:3000.
9. Pixel runs hourly reports via a cron job at https://pixel.wmcloud.org/. On the next run, Pixel will automatically pull down the latest database image without any manual intervention if everything is working correctly.