chore: clean up wof dictionary generation code

[missinglink]

I attempted to update the WOF resources today but lost enthusiasm half way through due to a bunch of different changes popping up.

This PR makes subsequent attempts at updating WOF considerably easier:

• update resources/whosonfirst/generate.js to sort dictionaries before writing them to disk (which makes diffs muuuch easier to read), also added some comments about potential gotchas
• sort existing dictionaries using the same sorting method (using the monstrous command below)

find resources/whosonfirst/dictionaries -type f -name '*.txt' \
  | node -e 'const fs=require(`fs`); fs.readFileSync(0, `utf-8`).trim().split(`\n`).forEach(file => fs.writeFileSync(file, fs.readFileSync(file, `utf-8`).trim().split(`\n`).sort().join(`\n`)))'

this should be a no-op, it's only sorting existing dictionaries, not adding or removing from them.

[missinglink]

😆

[Joxit]

Should we also normalize names here ? I found some name with uppercase and accents like Épinay even if LOWER is used in the SQLites statement (thank you French localities :sweat_smile:).

grep  'Épinay' resources/whosonfirst/dictionaries/locality/name\:fra_x_preferred.txt 

Second thought : Adding normalization will not improve diff reading, so I will merge as-is.