Reimplement delimit_words

[jgarte]

Hi,

Why does delimit_words need to be re-implemented?

You left a TODO comment here:

https://github.com/josiah-wolf-oberholtzer/uqbar/blob/main/uqbar/strings.py#L44

Is it because the function is too long?

[josiah-wolf-oberholtzer]

Probably. Probably could be mostly solved by a regex.

[jgarte]

What was the idea with the wordlike_characters checking variable?

I see in a repl that this function ignore characters like @, #, $, %, ^, &, *, etc.

>>> list(delimit_words("<hello><world>"))
['<', 'hello', '><', 'world', '>']

>>> list(delimit_words("!!!go go power rangers!!!"))
['!!!', 'go', 'go', 'power', 'rangers', '!!!']

>>> list(delimit_words("@parse;here@"))
['parse', 'here']

[josiah-wolf-oberholtzer]

Looks like it's mainly used to break up class names into component words, e.g. AbstractClass -> ["Abstract", "class"]

https://github.com/josiah-wolf-oberholtzer/uqbar/blob/2ec2309daca511117018bf14eb0d6b50938f50ca/uqbar/apis/graphs.py#L354

Are you looking to reimplement this, or just asking questions? If you're just poking around, take it to the discussions section you previously asked for.

[jgarte]

I would like to reimplement it, yes, for practice, but I'm first trying to scope out what would be accepted/needed for it in a PR.

Let me know if I should do that first over at discussions or just keep the thread here.

Feel free to leave the issue closed for now if it creates too much noise.

Looks like it's mainly used to break up class names into component words, e.g. AbstractClass -> ["Abstract", "class"]

I'll study that next, thanks!