Implement vale as pre-commit hook

.pre-commit-config.yaml

@@ -98,6 +98,12 @@ repos:
           - [email protected]
         args: ["--fix"]
 
+  - repo: https://github.com/errata-ai/vale
+    rev: 16d3a7f
+    hooks:
+      - id: vale
+        args: [--config=.vale/.vale.ini]
+
 # Configuration for https://pre-commit.ci/.
 ci:
   autoupdate_schedule: monthly

vizro-ai/docs/.vale/styles/Microsoft/Dashes.yml → .vale/styles/Microsoft/Dashes.yml

@@ -8,6 +8,6 @@ action:
   name: edit
   params:
     - remove
-    - ' '
+    - " "
 tokens:
   - '[—–]\s|\s[—–]'

.vale/styles/Microsoft/GenderBias.yml

@@ -0,0 +1,44 @@
+extends: substitution
+message: "Consider using '%s' instead of '%s'."
+ignorecase: true
+level: error
+swap:
+  (?:alumna|alumnus): graduate
+  (?:alumnae|alumni): graduates
+  air(?:m[ae]n|wom[ae]n): pilot(s)
+  anchor(?:m[ae]n|wom[ae]n): anchor(s)
+  authoress: author
+  camera(?:m[ae]n|wom[ae]n): camera operator(s)
+  chair(?:m[ae]n|wom[ae]n): chair(s)
+  congress(?:m[ae]n|wom[ae]n): member(s) of congress
+  door(?:m[ae]|wom[ae]n): concierge(s)
+  draft(?:m[ae]n|wom[ae]n): drafter(s)
+  fire(?:m[ae]n|wom[ae]n): firefighter(s)
+  fisher(?:m[ae]n|wom[ae]n): fisher(s)
+  fresh(?:m[ae]n|wom[ae]n): first-year student(s)
+  garbage(?:m[ae]n|wom[ae]n): waste collector(s)
+  lady lawyer: lawyer
+  ladylike: courteous
+  landlord: building manager
+  mail(?:m[ae]n|wom[ae]n): mail carriers
+  man and wife: husband and wife
+  man enough: strong enough
+  mankind: human kind
+  manmade: manufactured
+  manpower: personnel
+  men and girls: men and women
+  middle(?:m[ae]n|wom[ae]n): intermediary
+  news(?:m[ae]n|wom[ae]n): journalist(s)
+  ombuds(?:man|woman): ombuds
+  oneupmanship: upstaging
+  poetess: poet
+  police(?:m[ae]n|wom[ae]n): police officer(s)
+  repair(?:m[ae]n|wom[ae]n): technician(s)
+  sales(?:m[ae]n|wom[ae]n): salesperson or sales people
+  service(?:m[ae]n|wom[ae]n): soldier(s)
+  steward(?:ess)?: flight attendant
+  tribes(?:m[ae]n|wom[ae]n): tribe member(s)
+  waitress: waiter
+  woman doctor: doctor
+  woman scientist[s]?: scientist(s)
+  work(?:m[ae]n|wom[ae]n): worker(s)

vizro-core/docs/.vale/styles/Microsoft/HeadingPunctuation.yml → .vale/styles/Microsoft/HeadingPunctuation.yml

@@ -8,6 +8,6 @@ action:
   name: edit
   params:
     - remove
-    - '.!'
+    - ".!"
 tokens:
   - '[a-z][.!](?:\s|$)'

vizro-core/docs/.vale/styles/Microsoft/Headings.yml → .vale/styles/Microsoft/Headings.yml

@@ -5,7 +5,7 @@ level: suggestion
 scope: heading
 match: $sentence
 indicators:
-  - ':'
+  - ":"
 exceptions:
   - Kedro
   - Kedro-Viz

vizro-ai/docs/.vale/styles/Microsoft/Hyphens.yml → .vale/styles/Microsoft/Hyphens.yml

@@ -8,7 +8,7 @@ action:
   name: edit
   params:
     - replace
-    - '-'
-    - ' '
+    - "-"
+    - " "
 tokens:
   - '\s[^\s-]+ly-'

vizro-core/docs/.vale/styles/Microsoft/Negative.yml → .vale/styles/Microsoft/Negative.yml

@@ -7,7 +7,7 @@ action:
   name: edit
   params:
     - replace
-    - '-'
-    - '–'
+    - "-"
+    - "–"
 tokens:
   - '\s-\d+\s'

vizro-core/docs/.vale/styles/Microsoft/RangeTime.yml → .vale/styles/Microsoft/RangeTime.yml

@@ -7,7 +7,7 @@ action:
   name: edit
   params:
     - replace
-    - '[-–]'
-    - 'to'
+    - "[-–]"
+    - "to"
 tokens:
   - '\b(?:AM|PM)\s?[-–]\s?.+(?:AM|PM)\b'

vizro-ai/docs/.vale/styles/Microsoft/Semicolon.yml → .vale/styles/Microsoft/Semicolon.yml

@@ -5,4 +5,4 @@ nonword: true
 scope: sentence
 level: suggestion
 tokens:
-  - ';'
+  - ";"

vizro-ai/docs/.vale/styles/Microsoft/Spacing.yml → .vale/styles/Microsoft/Spacing.yml

@@ -4,4 +4,4 @@ link: https://docs.microsoft.com/en-us/style-guide/punctuation/periods
 level: error
 nonword: true
 tokens:
-  - '[a-z][.?!] {2,}[A-Z]'
+  - "[a-z][.?!] {2,}[A-Z]"

vizro-ai/docs/.vale/styles/Microsoft/Terms.yml → .vale/styles/Microsoft/Terms.yml

@@ -7,15 +7,15 @@ ignorecase: true
 action:
   name: replace
 swap:
-  '(?:agent|virtual assistant|intelligent personal assistant)': personal digital assistant
-  '(?:drive C:|drive C>|C: drive)': drive C
-  '(?:internet bot|web robot)s?': bot(s)
-  '(?:microsoft cloud|the cloud)': cloud
-  '(?:mobile|smart) ?phone': phone
-  '24/7': every day
-  'audio(?:-| )book': audiobook
-  'back(?:-| )light': backlight
-  'chat ?bots?': chatbot(s)
+  "(?:agent|virtual assistant|intelligent personal assistant)": personal digital assistant
+  "(?:drive C:|drive C>|C: drive)": drive C
+  "(?:internet bot|web robot)s?": bot(s)
+  "(?:microsoft cloud|the cloud)": cloud
+  "(?:mobile|smart) ?phone": phone
+  "24/7": every day
+  "audio(?:-| )book": audiobook
+  "back(?:-| )light": backlight
+  "chat ?bots?": chatbot(s)
   adaptor: adapter
   administrate: administer
   afterwards: afterward
@@ -29,7 +29,7 @@ swap:
   anti-virus: antivirus
   appendixes: appendices
   artificial intelligence: AI
-  '(?:assembler|machine language)': assembly language
+  "(?:assembler|machine language)": assembly language
   caap: CaaP
   conversation-as-a-platform: conversation as a platform
   eb: EB

vizro-ai/docs/.vale/styles/Microsoft/Units.yml → .vale/styles/Microsoft/Units.yml

@@ -5,10 +5,10 @@ level: error
 raw:
   - '[a-zA-Z]+\s'
 tokens:
-  - '(?:centi|milli)?meters'
-  - '(?:kilo)?grams'
-  - '(?:kilo)?meters'
-  - '(?:mega)?pixels'
+  - "(?:centi|milli)?meters"
+  - "(?:kilo)?grams"
+  - "(?:kilo)?meters"
+  - "(?:mega)?pixels"
   - cm
   - inches
   - lb

vizro-ai/docs/.vale/styles/Microsoft/Vocab.yml → .vale/styles/Microsoft/Vocab.yml

@@ -1,6 +1,6 @@
 extends: existence
 message: "Don't use '%s'!"
-link: 'https://docs.microsoft.com/en-us/style-guide'
+link: "https://docs.microsoft.com/en-us/style-guide"
 level: error
 ignorecase: true
 tokens:

.vale/styles/Microsoft/links.yml

@@ -0,0 +1,23 @@
+extends: substitution
+message: "Avoid vague text in links like '%s' unless you can pair it with more descriptive text."
+ignorecase: true
+scope: raw
+nonword: true
+level: error
+
+swap:
+  # For the word 'here' in Markdown and HTML links
+  '\[here\]\(.*?\)': "here"
+  '<a\s*href\s*=\s*".*?".*?>\s*here\s*</a>': "here"
+
+  # For the word 'this' in Markdown and HTML links
+  '\[this\]\(.*?\)': "this"
+  '<a\s*href\s*=\s*".*?".*?>\s*this\s*</a>': "this"
+
+  # For the word 'page' in Markdown and HTML links
+  '\[page\]\(.*?\)': "page"
+  '<a\s*href\s*=\s*".*?".*?>\s*page\s*</a>': "page"
+
+  # For the phrase 'this page' in Markdown and HTML links
+  '\[this page\]\(.*?\)': "this page"
+  '<a\s*href\s*=\s*".*?".*?>\s*this page\s*</a>': "this page"