The main goal of this PR is to update the 1wg-charters.txt and 1wg-charters-by-acronym.txt files via celery tasks instead of cron jobs. The cron job worked by using wget to retrieve a datatracker URL that generated the files. This was done once per hour and cached for an hour each time using the slowpages cache.

As a result, the slowpages cache is effectively always holding on to the data generated by the last run of bin/hourly and the view itself is always served from that cache except when the cron job regenerates it.

This seems needlessly baroque, so I've refactored the work at generating the charters files entirely out of the views and into a task. The views then simply grab the file the task updates and hand that back in their response.

The new method has a bunch of benefits. It avoids collecting the group data twice (the views gathered the same data twice, the task does it only once). It avoids the small chance that the charter data is occasionally an extra hour out of date depending on system load (the cron job might make its request at very slightly less than an hour since the previously cached value). It entirely moves the purportedly expensive data gathering out of the view.

Important: I did a careful comparison of the output of the old and new code and there are a couple of differences. Most notably, the groups in 1wg-charters.txt are sorted alphabetically by acronym within each area. Before, they were (to my eye) random within the area. This seems like an improvement to me, but it is a change.

Second, there were some \r\n line endings in the old output - I assume they were bleeding in from the charters. These are converted to \n by the new rendering path. This seems like an absolute win to me, but again it's a change.

Also important: We reviewed the \r (aka CR) changes and found it was caused by CRLFs in some old draft titles, not in the charters. Added a clean_whitespace filter to the title in the template to fix this. That replaces any run of ASCII whitespace / control characters (i.e., 0x00-0x1F) with a single space, then strips leading and trailing space. This ends up improving spacing / inappropriate line breaks in a couple dozen draft titles.