Get register of interest via API

[ajparsons]

Parliament's turned off their old site, so our scraper is broken!

This scopes out the old parser, adds a new parser that converts the Parliament API into the XML TheyWorkForYou expects.

As seen in dailyupdate, this can be run via:

poetry run python -m pyscraper.regmem download-all-registers --chamber commons

This updates some requirements in the poetry, and also uses this poetry config for the first time in production.

Rather than exactly replicating the previous structure, there's now a bit more use of lists to reflect how parliament sees items and subitems.

I've tested this in both the import to database:

And the comparison over time:

There's a boundary problem in that the sudden change will break the retrospect. If we delete and recreate the last few items though, this should be less of a problem.

Also have the option if reviewing this will take a while to just upload the generated files manually (given they seem fine).

[ajparsons]

twfy host now has poetry - so this should work even in advance of the twfy poetry pr.

[dracos]

One issue with missing jinja2. To be honest, I didn't follow all the code (perhaps some doctests of e.g. move_subitems_under_parent to show an example would be nice, not sure I understood the "parents" stuff), but if you're happy with the output, fine with me :)

[ajparsons]

Added above - added some more commentary on the subitem movements because that wasn't clear.

Added one extra fix that was needed for a march register to manually fetch a not present parent.

[dracos]

Thanks!