Update code to work with Cliopatria "composite" shapes (04062024)

[edwardchalstrey1]

• Handle shapes linked to multiple polities separated by semi-colon #108
  • Latest version seems to lack this
• A bunch of the shapes in Cliopatria.geojson have a question mark in the SeshatID field, which is causing problems with linking - it doesn't really matter, but you could update the populate script to ignore these (actually we still want these)
• Polities with the same Seshat ID should have the same colour
• (British Empire) still has brackets
• Polities including Qing dynasty missing
• Error: Uncaught (in promise) SyntaxError: Unexpected end of JSON input (is this down to there being too many shapes?
• Why is it so much slower now?
  • There are now 38,411 shapes where before there was 10,310 - check why this is and ask Jim if unsure
  • The number of polities has slightly gone down from 1,346 to 1,325 so that's not the issue
  • Loads 3-5 times slower
  • Edit: Jim said he made a mistake: "Turns out the drop I sent you was a ‘full’ drop made without running the code to drop polygons for polities where they are duplicated"
• Locally load the previous Cliopatria for now, which will ensure the script is backwards compatible. Make a separate db so you have both

Info:

From Jim Bennett:

Note this version has the ‘composite’ shapes with names that are surrounded by ()’s.
Ed, this may mess things up for you. In particular you will get both ‘(Roman Empire)’ and ‘Roman Empire’. The later is the low-level shape; you’ll want to display the (Roman Empire) composite versions instead. This might require a two-pass alg, the first to find those names and them drop the equivalent unadorned versions on the internment pass. You might also want to strip the leading and trailing parens. (Note that there are sometimes internal parens, e.g., our old friend ‘Peithon (son of Agenor)’ so just don’t .replace(‘(‘,’’).replace(‘)’,’’)!)

[edwardchalstrey1]

I think what he means is that there will be some "low level" shapes that belong to multiple polities, but doing it this way avoids the #108 situation. The result will mean that some of the low level shapes will be represented twice in the same year, resulting in an overlap of polities - which is actually good because that makes sense and is what we're trying to represent by saying a low level shape (a region) is part of both polities

[edwardchalstrey1]

Closed by #114