use regex for lexical illusions

[Nytelife26]

This relates to...

The usage of regular expressions rather than simple matches for
lexical_illusions.misc.

Rationale

Prior to this, the lexical illusions check would only catch
3 different lexical illusions, and always in pairs of two rather than
flagging the entire strand. This is inefficient, and does not catch
many lexical illusions at all.

Changes

• lexical_illusions.misc now uses regex
• lexical_illusions.misc now reports entire strands instead of many word pairs
• lexical_illusions.misc reports all lexical illusions except for that that
  and had had

Features

N/A.

Bug Fixes

N/A.

Breaking Changes and Deprecations

N/A.

[codecov]

Codecov Report

Merging #1174 (b82f8bb) into main (8b3498c) will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##             main    #1174   +/-   ##
=======================================
  Coverage   90.14%   90.15%           
=======================================
  Files          83       83           
  Lines        1208     1209    +1     
=======================================
+ Hits         1089     1090    +1     
  Misses        119      119           

Flag	Coverage Δ
macos-latest	90.15% <100.00%> (+<0.01%)	⬆️
py3.6	89.19% <100.00%> (+<0.01%)	⬆️
py3.7	89.19% <100.00%> (+<0.01%)	⬆️
py3.8	90.15% <100.00%> (+<0.01%)	⬆️
py3.9	90.15% <100.00%> (+<0.01%)	⬆️
pypypy3	89.19% <100.00%> (+<0.01%)	⬆️
ubuntu-latest	90.15% <100.00%> (+<0.01%)	⬆️
windows-latest	90.15% <100.00%> (+<0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
proselint/checks/lexical_illusions/misc.py	100.00% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 8b3498c...b82f8bb. Read the comment docs.

[suchow]

There are a few instances where repeated words are perfectly acceptable. For example, one day I was commenting on student writing and I noticed that the student used the word "that" instead of "which". It's a hard distinction for many, and I don't blame them. So here I am now, telling you that that "that" that that student wrote ought to have been a "which".

[suchow]

See also https://en.wikipedia.org/wiki/Buffalo_buffalo_Buffalo_buffalo_buffalo_buffalo_Buffalo_buffalo

[suchow]

https://en.wikipedia.org/wiki/James_while_John_had_had_had_had_had_had_had_had_had_had_had_a_better_effect_on_the_teacher

[Nytelife26]

There are a few instances where repeated words are perfectly acceptable.

Gramatically valid, perhaps. Stylistically speaking, it's an absolute atrocity. I had to attempt to read that example you threw in around 3 times before it parsed correctly in my mind. Valid, yes - but also beneficial (and always possible) to avoid. An exception has been made for punctuation inbetween word boundaries, because splitting the lexical structure into phrases makes it readable. I can therefore tell you that the "that" that the student wrote should have been a "which", but there are other ways of writing about the "which" which the student should've written without entering linguistic territory in which a "which" which your student should've written such that the "that" that your student wrote isn't so bad.

[Nytelife26]

https://en.wikipedia.org/wiki/James_while_John_had_had_had_had_had_had_had_had_had_had_had_a_better_effect_on_the_teacher

"is an English sentence used to demonstrate lexical ambiguity and the necessity of punctuation" I do not think from a stylistic perspective we should be referencing esoteric sentences that serve to demonstrate poor linguistic usage when justifying limitations. ``` The sentence is easier to understand with added punctuation and emphasis: James, while John had had "had", had had "had had"; "had had" had had a better effect on the teacher. ``` The tool is designed to help people write better. Avoiding structures like the ones you have referenced in favour of their better formatted alternatives is doing exactly that, I feel. Either way, the decision lies with you. I've said my piece.

[suchow]

How about we make an exception for that that (two in a row, not more) and for had had (which is the past perfect form of the verb to have and hardly a verbal atrocity, e.g., "He had had one too many drinks."), and you can have the rest? The links were meant to be fun examples, not things you needed to take seriously, which is why I left them as comments and didn't mark the changes as necessary in the review.

[Nytelife26]

How about we make an exception for that that (two in a row, not more) and for had had (which is the past perfect form of the verb to have and hardly a verbal atrocity, e.g., "He had had one too many drinks."), and you can have the rest?

Works for me. Apologies if that came off aggressive, by the way, I was just sharing my thoughts and it ended up coming out somewhat passionately

[suchow]

All good, not going to complain about passion!

[Nytelife26]

It seems like to do so we will need to be able to create exceptions using the
existence check. Unfortunately, Regex does not present us with a viable way to
do this (no amount of negative lookaround hacks would work), so we may need to
use a separate PR to implement that - quite a breaking change.

[Nytelife26]

@suchow The exceptions for had had and that that have been successfully implemented. Furthermore, the lexical illusions check will now flag the entire strand of lexical illusions, rather than repeatedly flagging it in word pairs of two - much more efficient.

[suchow]

Wonderful, thank you, an excellent way to resolve this issue.