Reorgansing data in `FluffyChatBotDictionaries.py`

[aneroid]

Might be easier to maintain data in separate text-based files instead of large sets and dictionaries all in one big Python file.

• FluffyChatBotDictionaries.py is now a module with the same name, and the data is loaded/imported in __init__.py.
  • So the from FluffyChatBotDictionaries import UnitNameDict, CommanderMastery, ... line in ReplayAnalysis.py needs no changes. And the referenced data structures are the same as the original.
• Moved dictionaries to .csv files and sets to .txt, while maintaining order. (Shouldn't matter for sets though.)
  • csv's have headers for human readability but are ignored when loaded.
  • txt's have no headers and are just meant to be loaded into sets, lists, etc.
  • comastery dicts are specifically in a key: [value1, ..., value6] format.
• Added functions in _data_utils.py which are used in FCB.init to load them appropriately
  • there is an optional factory parameter when loading txt's which defaults to set as per the original data
  • the param can be set to list, frozenset, etc. as needed
  • empty lines in either format are ignored
• All the local names and imports are _underscored so as not to pollute the namespace with local vars or the helper functions.
• There is some size reduction since the raw data doesn't require Python delimiters.
  • Example: UnitNames raw dict was 48k characters. The UnitNames.csv is 41 KB. (In-memory size remains the same of course.)

[aneroid]

That just leaves UnitCompDict in the init file since that's a more complex structure - dict values are each a list of sets. I could move that out to a csv and add a loader specifically for it but it seems like more effort than it's worth, especially when you want to edit/maintain that data in its text-based format.

I'd recommend just moving it out to its own .py file and importing it (in init). Since that change is a bit iffy, I've left it out of this pull request.

Let me know if it makes sense to you to have it that way. Would take all of 5 mins to do it, and transparently so that the import in ReplayAnalysis.py remains unchanged.

[thendarion]

I like the idea from over here (ie. without any stakes in the matter)

though I would totally squash those commits into one

[aneroid]

Thanks!

I had originally done it as just 2 bulk changes. Decided to break them down per data item so that once merged, FluffyMaguro could easily revert an item if he/she/they decided that it was better left in FCBD.__init__.py. But as a bulk change, it would need to be done afterwards and manually.

As a branch, if reorg_data is merged with --no-ff (no fast forward), it keeps the main branch clean. In any case, they could still choose to merge it --squashed.

[FluffyMaguro]

Thank you, I'll add that.