Brian Skinn
Brian Skinn
Can consider using https://pypi.org/project/editdistance/ as a new `speedup` extra, integrating into the new `default` scorer. Or --- if it's fast enough with the `editdistance` speedup, may be able to avoid...
Sprinkling this in various issues: [`rapidfuzz`](https://pypi.org/project/rapidfuzz/) as another `fuzzywuzzy` alternative.
I don't think this change to the inventory semantics is necessary. The [most recent regex] I implemented for `sphobjinv`, which incorporated a couple of tweaks from https://github.com/bskinn/sphobjinv/issues/181 and https://github.com/bskinn/sphobjinv/issues/256, has...
> but there are some titles such as `word word1:word2 integer` that would fail your regex since `word1:word2` would be incorrectly matched as `domain:reftype`. Indeed, that is a point of...
Definitely true, including the name length as you propose is probably about the most minimal change achievable, and wouldn't dramatically change the flow of line parsing. --- Quick size analysis,...
Re your regex, I think you have a typo in the first line: ```none (?P(?:.|[^\s:]+:\S+\s+-?\d+\s*)+)? ``` Is that trailing `?` supposed to be inside that final `)`? Where it's located...
> If we include the title length, I think we'll probably have less Yes, agreed, yours is much more concise. > because, who in their right mind would have a...
> I'd be glad to have it. It could help a lot for [#12204](https://github.com/orgs/sphinx-doc/discussions/12204) Ok, here's what I turned up, in no particular order: - Possible colon in role: -...
However: Since the regex pattern compilation at this point has two flags that change the capture-group insertion behavior, pre-compilation of the patterns *may* be more convoluted than would be useful.
Oh, *interesting* -- I didn't realize that was a limitation of the INI format. Yes, it makes total sense! Would it make sense to add a warning, displayed to the...