Handle newlines in the middle of a name

[VirginiaDooley]

co-authored with @Bekabyx

Closes #2106

While my initial fix was more of any improvement for this specific SOPN, in the process of review, @Bekabyx discovered it was actually camelot that is responsible for adding new lines. At her suggestion, I've moved the code to strip newlines from the SOPN earlier in the process, (we don't store addresses so as long as this is the case, this change does not post any issues, AFAIK) into the extract table step. I've updated tests to reflect this change to the process.

Test locally by uploading the SOPN for https://candidates.democracyclub.org.uk/bulk_adding/sopn/local.west-lancashire.rural-south.2023-06-22/ and parse. Names should be formatted as expected.

[VirginiaDooley]

Handle newlines in the middle of a name

[Bekabyx]

This is working as expected on my local machine, SOPN is parsing correctly and all names + party descriptions look fine. As long as we're cool with not having newlines in addresses I'm happy for this to merge 😁

[VirginiaDooley]

Tattenhams Residents' Association

[VirginiaDooley]

@symroe Based on SOPN testing (updated results in Trello card), there was a decrease in people parsed as a result of this change. I think we should not merge this work and instead focus parsing improvements in conjunction with the spike into AWS textract.

[VirginiaDooley]

Closing based on #2107 (comment) and opening https://trello.com/b/N1TKOKQ5/dc-icebox?search_id=e2ecd086-9537-40f5-bac0-53a66235b90f. This solution proved to be too nuclear and led to some missed matches with parties. It's still worthwhile knowing about these issues so we can use them in testing in the spike.