Run scraper with --stats once per week

[simonw]

This is so I don't get lots of tiny diffs because of page view and download counts incrementing all the time.

I built the script with this in mind - it only writes the stats information out - as separate files - if you include --stats:

scrape-open-data/scrape_socrata.py

Lines 28 to 31 in 626c4cb

	if save_stats:
	if domain not in stats_files:
	stats_files[domain] = (root / "{}.stats.json".format(domain)).open("w")
	stats_files[domain].write(json.dumps(stats) + "\n")

[simonw]

I can use this pattern: https://til.simonwillison.net/github-actions/different-steps-on-a-schedule

[simonw]

Note that with this change the action no longer scrapes on a commit - it only scrapes on workflow_dispatch or when the schedules trigger.

[simonw]

Running now with workflow_dispatch which should populate the stats files for the first time.

[simonw]

Yup, that added the stats files: 1a09c87

[simonw]

I manually ran it again to check I got some diffs and I did: 2060f38#diff-6835345cbfec8fbf1dfeaee6534859a57591cf163fae958e06defcc40f87b969