workflow: move all actions off the Node 20 runtime

GitHub is removing the Node 20 action runtime on September 23rd, 2026. Every action used here still runs on it, and actions/checkout was pinned to v3, which is on the already-retired Node 16 runtime.

Bump each action to the current major, all of which run on Node 24:

checkout v3 -> v7 setup-python v5 -> v7 upload-artifact v4 -> v7 configure-pages v5 -> v6 upload-pages-artifact v3 -> v5 deploy-pages v4 -> v5 download-artifact v5 -> v8 setup-node v4 -> v7

upload-pages-artifact is a composite action; v5 pins upload-artifact v7 internally, so it is off Node 20 as well. Every input passed here still exists in the new majors, and node-version remains optional on setup-node, so the bare “Set up Node.js” step keeps using the runner’s preinstalled Node.

Co-Authored-By: Claude Opus 5 (1M context) noreply@anthropic.com Claude-Session: https://claude.ai/code/session_014XsRepKWSse3HXAqNbM9Lm

PR Dashboard

This set of scripts produces a static HTML page with an overview of the pull requests involving a specific user, fetching data from multiple repositories under a GitHub organization. It’s been developed for the Zephyr project specifically but it should be usable for any GitHub project.

Script architecture

This is meant to be called periodically from a GitHub workflow and the output served using GitHub pages. The basic workflow run as following:

update_zephyr_pr.py: produces a list of repositories based on the west manifest and calls update_pr.py with the list. This step is optional, one could call update_pr.py with a static list of repositories directly.
update_pr.py: fetches the pull request data using the GitHub APIs and dump the raw content into a big cache file that can be used by other scripts to read the data without having to query the GitHub APIs again.
crunch_data.py: reads the raw data and produces a new set of files that are suitable for the dashboard UI to be used directly.

See the .github/workflows/update.yaml file for more details.

Development and troubleshooting

Each of the scripts can run independently.

update_pr.py can be called directly and only needs a GITHUB_TOKEN environment variable to be setup with a valid API token. By default it only fetches a couple of fixed repositories.
crunch_data.py can be run using the cached data from the latest data set already served on GitHub, these can be fetched using the download_cache_data script:

~/pr-dashboard$ mkdir cache
~/pr-dashboard$ ./download_cache_data 
...
2025-01-07 12:12:52 (18.0 MB/s) - ‘cache/data_dump.json.bz2’ saved [96004/96004]

~/pr-dashboard$ ls cache/
data_dump.json
~/pr-dashboard$ ./crunch_data.py
...
~/pr-dashboard$ ls public/
index.html  prs.json  style.css  users.json
~/pr-dashboard$ cd public/
~/pr-dashboard/public$ python -m http.server
Serving HTTP on 0.0.0.0 port 8000 (http://0.0.0.0:8000/) ...