Case ignore

[vEnhance]

Starting on #1578, WIP

[DimitriPapadopoulos]

#3270 makes all typos (but not suggested fixes) lowercase. Does your implantation require #3270 as a prior?

[vEnhance]

#3270 makes all typos (but not suggested fixes) lowercase. Does your implantation require #3270 as a prior?

It should be independent. The implementation of this PR only affects the reading of -I and -L options and leaves everything else the same.

In fact, the build_dict function that parses the built-in dictionaries currently reads (this is not changed by this PR):

def build_dict(
    filename: str,
    misspellings: Dict[str, Misspelling],
    ignore_words: Set[str],
) -> None:
    with open(filename, encoding="utf-8") as f:
        translate_tables = [(x, str.maketrans(x, y)) for x, y in alt_chars]
        for line in f:
            [key, data] = line.split("->")
            # TODO for now, convert both to lower. Someday we can maybe add
            # support for fixing caps.
            key = key.lower()
            data = data.lower()
            if key not in ignore_words:
                add_misspelling(key, data, misspellings)
            # generate alternative misspellings/fixes
            for x, table in translate_tables:
                if x in key:
                    alt_key = key.translate(table)
                    alt_data = data.translate(table)
                    if alt_key not in ignore_words:
                        add_misspelling(alt_key, alt_data, misspellings)

that is, when reading the dictionaries included with codespell, actually everything is already converted into lowercase anyway.

[DimitriPapadopoulos]

@larsoner Would you like to have a look?

[larsoner]

Code changes look reasonable. There is some risk that someone with custom dictionaries will get different behavior with this PR, though. Safest would be to add a switch to make this opt-in like --ignore-words-case-sensitive or similar. Less safe but maybe useful (or maybe YAGNI) would be to add a --ignore-words-case-insensitive that could get back the old behavior, but people can also get the old behavior just by modifying their dictionaries. I think I can live with the backward incompatible change but wanted to make sure we thought about it first!

[vEnhance]

There is some risk that someone with custom dictionaries will get different behavior with this PR, though.

So I believe the changes in this PR should only affect words passed by config or command line options, which is why I stated in #3272 (comment) this was independent of #3270. Similarly, there should be no backwards-incompatability with custom dictionaries either, because those are read using build_dict which has remained unchanged.

The only users who should see any changes are people who have specified, say, something like SEH in -I which, as mentioned previously (e.g. #1578 (comment)), actually did nothing. I can't actually imagine any situations where someone would add a word into an ignore list and want/expect it to have no effect.

Please let me know if I misunderstood.

[larsoner]

Thanks @vEnhance !