errata-ai/write-good

A Vale-compatible implementation of the write-good linter.

write-good's Issues

Remove the token `only` from Weasel words

btford/write-good#92 indicates it is not a weasel word and should be removed from here:

write-good/write-good/Weasel.yml

Line 127 in 2d11661

- only

I would wish to push a commit to fix this if it is ok.

Slightly related: #9.

E-Prime false positive: be

Try to avoid using 'be'. is suggested for e.g. before and best.

Update

Is it possible for a new release of this implementation closer to write-good's last update?

write-good.ThereIs

The sentence This is the where there are cows. triggers an error Don't start a sentence with 'there are'.

"It's" missing from E-Prime?

I can find "is" and "isn't" in E-Prime:

write-good/write-good/E-Prime.yml

Lines 17 to 18 in 769e957

	- is
	- isn't

but I don't see "it's". Shouldn't it be there?

How do I ignore an instance of TooWordy?

I've written "Set the minimum version for each dev dependency."

Vale warns that "minimum" is too wordy.

In this technical context I think it's the right word to use.

Can I tell Vale to ignore just this instance of the word (say by line number) and warn me anywhere else?

Illusions incorrectly flags lines

Given the following rst:

foo foo

foo bar

foo ``bar`` foo bar

foo ``bar`` foo ``bar``

foo ``bar`` foo ``bar`` cows

The output is:

 1:1  warning  'foo' is repeated!  write-good.Illusions 
 5:1  warning  'foo' is repeated!  write-good.Illusions 
 7:1  warning  'foo' is repeated!  write-good.Illusions 
 9:1  warning  'foo' is repeated!  write-good.Illusions 

Expected output should only contain line 1. My guess is that vale removes the literal (the text between double backticks) before applying the regex.

Note that if you remove the empty lines, the output is different even though it should be the same.

"also" is not "so"

There may be a false positive for the word "also" in So.yml. The file looks for "so followed by a space" but this is also true when the word also is followed by a space. Perhaps, instead, it should look for "so preceded by a period and/or space(s)" and also look for "so preceded by line break(s)"

Unfortunately I can read regex fairly well, but I'm unsure how to write it, or else I'd try my hand at a pull request.

write-good rules are triggered for words detected within compound words

❯ vale --version
vale version "2.20.1"

vale ls-config:

{
  "BlockIgnores": {
    "*.md": [
      "(?sm)^({{[%\u003c] [^{]*? [%\u003e]}})\\n$",
      "(?s) *({{\u003c highlight [^\u003e]* ?\u003e}}.*?{{\u003c ?/ ?highlight \u003e}})"
    ]
  },
  "Checks": null,
  "Formats": {
    "ex": "md",
    "exs": "md"
  },
  "Asciidoctor": {},
  "FormatToLang": {},
  "GBaseStyles": null,
  "GChecks": {},
  "IgnoredClasses": null,
  "IgnoredScopes": null,
  "MinAlertLevel": 0,
  "Vocab": [
    "base"
  ],
  "RuleToLevel": {},
  "SBaseStyles": {
    "*.{txt,org,md}": [
      "Vale",
      "write-good",
      "proselint",
      "alex",
      "Joblint"
    ]
  },
  "SChecks": {
    "*.md": {},
    "*.{txt,org,md}": {}
  },
  "SkippedScopes": null,
  "Stylesheets": {},
  "StylesPath": "/home/john/.config/vale/styles",
  "TokenIgnores": {
    "*.md": [
      "({{[%\u003c] .* [%\u003e]}}.*?{{[%\u003c] ?/.* [%\u003e]}})",
      "(\\[.+\\]\\({{\u003c .+ \u003e}}\\))"
    ]
  },
  "WordTemplate": "",
  "DictionaryPath": "",
  "NLPEndpoint": ""
}

❯ echo "And cares are different from yours. And yes! I made sure you knew it," | vale

 stdin.txt
 1:11  suggestion  Try to avoid using 'are'.  write-good.E-Prime 

✔ 0 errors, 0 warnings and 1 suggestion in stdin.

❯ echo "And snares are different from yours. And yes! I made sure you knew it," | vale

 stdin.txt
 1:12  suggestion  Try to avoid using 'are'.  write-good.E-Prime 

✔ 0 errors, 0 warnings and 1 suggestion in stdin.