buildtables overwrites files on (case-insensitive?) category collision

[ducky64]

An example is with inductors, which have category entries

id	category	subcategory
12	Inductors/Coils/Transformers	Inductors (SMD)
695	inductors/coils/transformers	Inductors (SMD)

On Windows at least, both write to InductorsakaCoilsakaTransformersInductors__SMD_.json.gz, so in practice the latter one overwrites the former. The latter one is also completely useless, containing just one part.

There's a bunch more of these cases.

Quick and dirty way to not get it to overwrite: in jlcparts/datatables.py, add:

  def saveJson(object, filename, hash=False, pretty=False, compress=False):
+     if os.path.isfile(filename):
+         print(f"**** dropping {filename}")
+         return

A cleaner solution may be to append the tables instead of overwriting, treating it as one giant file. Which doesn't adhere to the input data as strictly (especially in terms of category ID), but maybe is closer to the spirit of the data?

Actually, category name standardization (across case, / vs. & vs. , ) could probably clean up the data quite a bit