ChatGPT preprocessing - make longer ordinances work

[maral]

Currently the script is pretty reliable for cities with up to 6 schools but the larger cities get stuck. This can be solved in two ways:

1. Keep the prompt as-is and if the resulting JSON is incomplete, add another prompt that would tell the API to continue where it left off. E.g.:
   Skončil jsi v polovině školy. Prosím, pokračuj přesně tam, kde jsi skončil, zopakuj poslední řádek, který jsi napsal a pokračuj dál. Opět pouze JSON.
   Then you'd have to connnect the 2+ outputs.

2. Another strategy is to first let ChatGPT extract only the school names and in the follow-up prompts let it return only several schools per script - this could be also much faster, as you can do multiple calls in parallel. Then you just need to connect the results together. This is probably the better solution, but a slightly more work.