# See https://github.com/check-spelling/check-spelling/wiki/Configuration-Examples:-patterns

# Automatically suggested patterns

# hit-count: 654 file-count: 86
# https/http/file urls
(?:\b(?:https?|ftp|file)://)[-A-Za-z0-9+&@#/%?=~_|!:,.;]+[-A-Za-z0-9+&@#/%=~_|]

# hit-count: 98 file-count: 34
# scala imports
^import (?:[\w.]|\{\w*?(?:,\s*(?:\w*|\*))+\})+

# hit-count: 23 file-count: 6
# version suffix <word>v#
(?:(?<=[A-Z]{2})V|(?<=[a-z]{2}|[A-Z]{2})v)\d+(?:\b|(?=[a-zA-Z_]))

# hit-count: 19 file-count: 10
# Python string prefix / binary prefix
# Note that there's a high false positive rate, remove the `?=` and search for the regex to see if the matches seem like reasonable strings
(?<!['"])\b(?:B|BR|Br|F|FR|Fr|R|RB|RF|Rb|Rf|U|UR|Ur|b|bR|br|f|fR|fr|r|rB|rF|rb|rf|u|uR|ur)['"](?=[A-Z]{3,}|[A-Z][a-z]{2,}|[a-z]{3,})

# hit-count: 19 file-count: 2
# kubernetes object suffix
-[0-9a-f]{10}-\w{5}\s

# hit-count: 14 file-count: 13
# Contributor
\[[^\]]+\]\(https://github\.com/[^/\s"]+/?\)

# hit-count: 13 file-count: 5
# Compiler flags (Windows / PowerShell)
# This is a subset of the more general compiler flags pattern.
# It avoids matching `-Path` to prevent it from being treated as `ath`
(?:^|[\t ,"'`=(])-(?:[DPL](?=[A-Z]{2,})|[WXYlf](?=[A-Z]{2,}|[A-Z][a-z]|[a-z]{2,}))

# hit-count: 11 file-count: 10
# hex digits including css/html color classes:
(?:[\\0][xX]|\\u|[uU]\+|#x?|%23)[0-9_a-fA-FgGrR]*?[a-fA-FgGrR]{2,}[0-9_a-fA-FgGrR]*(?:[uUlL]{0,3}|[iu]\d+)\b

# hit-count: 9 file-count: 4
# node packages
(["'])@[^/'" ]+/[^/'" ]+\g{-1}

# hit-count: 8 file-count: 4
# libraries
\blib(?!rar(?:ies|y))(?=[a-z])

# hit-count: 8 file-count: 3
# AWS VPC
vpc-\w+

# hit-count: 7 file-count: 7
# uuid:
\b[0-9a-fA-F]{8}-(?:[0-9a-fA-F]{4}-){3}[0-9a-fA-F]{12}\b

# hit-count: 7 file-count: 7
# Compiler flags (Unix, Java/Scala)
# Use if you have things like `-Pdocker` and want to treat them as `docker`
(?:^|[\t ,>"'`=(])-(?:(?:J-|)[DPWXY]|[Ll]|f(?!ix))(?=[A-Z]{2,}|[A-Z][a-z]|[a-z]{2,})

# hit-count: 7 file-count: 7
# set arguments
\b(?:bash|sh|set)(?:\s+-[abefimouxE]{1,2})*\s+-[abefimouxE]{3,}(?:\s+-[abefimouxE]+)*

# hit-count: 7 file-count: 5
# hex runs
\b[0-9a-fA-F]{16,}\b

# hit-count: 4 file-count: 2
# kubernetes pod status lists
# https://kubernetes.io/docs/concepts/workloads/pods/pod-lifecycle/#pod-phase
\w+(?:-\w+)+\s+\d+/\d+\s+(?:Running|Pending|Succeeded|Failed|Unknown)\s+

# hit-count: 1 file-count: 1
# integrity
integrity=(['"])(?:\s*sha\d+-[-a-zA-Z=;:/0-9+]{40,})+\g{-1}

# hit-count: 1 file-count: 1
# curl arguments
\b(?:\\n|)curl(?:\.exe|)(?:\s+-[a-zA-Z]{1,2}\b)*(?:\s+-[a-zA-Z]{3,})(?:\s+-[a-zA-Z]+)*

# pom.xml id fields
<\w+Id>[^<]+</

# Questionably acceptable forms of `in to`
# Personally, I prefer `log into`, but people object
# https://www.tprteaching.com/log-into-log-in-to-login/
\b(?:[Ll]og|[Ss]ign) in to\b

# to opt in
\bto opt in\b

# acceptable duplicates
# ls directory listings
[-bcdlpsw](?:[-r][-w][-SsTtx]){3}[\.+*]?\s+\d+\s+\S+\s+\S+\s+\d+\s+
# mount
\bmount\s+-t\s+(\w+)\s+\g{-1}\b
# C types and repeated CSS values
\s(auto|center|div|inherit|long|LONG|none|normal|solid|thin|transparent|very)(?: \g{-1})+\s
# C struct
\bstruct\s+(\w+)\s+\g{-1}\b
# go templates
\s(\w+)\s+\g{-1}\s+\`(?:graphql|inject|json|yaml):
# doxygen / javadoc / .net
(?:[\\@](?:brief|groupname|link|t?param|return|retval)|(?:public|private|\[Parameter(?:\(.+\)|)\])(?:\s+(?:static|override|readonly|required|virtual))*)(?:\s+\{\w+\}|)\s+(\w+)\s+\g{-1}\s

# Python package registry has incorrect spelling for macOS / Mac OS X
"Operating System :: MacOS :: MacOS X"

# Commit message -- Signed-off-by and friends
^\s*(?:(?:Based-on-patch|Co-authored|Helped|Mentored|Reported|Reviewed|Signed-off)-by|Thanks-to): (?:[^<]*<[^>]*>|[^<]*)\s*$

# Autogenerated revert commit message
^This reverts commit [0-9a-f]{40}\.$

# ignore long runs of a single character:
\b([A-Za-z])\g{-1}{3,}\b

# hit-count: 40 file-count: 8
# Non-English
[a-zA-Z]*[ÀÁÂÃÄÅÆČÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖØÙÚÛÜÝßàáâãäåæčçèéêëìíîïðñòóôõöøùúûüýÿĀāŁłŃńŅņŒœŚśŠšŜŝŸŽžź][a-zA-Z]{3}[a-zA-ZÀÁÂÃÄÅÆČÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖØÙÚÛÜÝßàáâãäåæčçèéêëìíîïðñòóôõöøùúûüýÿĀāŁłŃńŅņŒœŚśŠšŜŝŸŽžź]*|[a-zA-Z]{3,}[ÀÁÂÃÄÅÆČÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖØÙÚÛÜÝßàáâãäåæčçèéêëìíîïðñòóôõöøùúûüýÿĀāŁłŃńŅņŒœŚśŠšŜŝŸŽžź]|[ÀÁÂÃÄÅÆČÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖØÙÚÛÜÝßàáâãäåæčçèéêëìíîïðñòóôõöøùúûüýÿĀāŁłŃńŅņŒœŚśŠšŜŝŸŽžź][a-zA-Z]{3,}

# hit-count: 22 file-count: 1
# ANSI color codes
(?:\\(?:u00|x)1[Bb]|\x1b|\\u\{1[Bb]\})\[\d+(?:;\d+|)m

# hit-count: 15 file-count: 3
# base64 encoded content
([`'"])[-a-zA-Z=;:/0-9+]{3,}=\g{-1}

# hit-count: 11 file-count: 2
# c99 hex digits (not the full format, just one I've seen)
0x[0-9a-fA-F](?:\.[0-9a-fA-F]*|)[pP]

# hit-count: 10 file-count: 8
# IServiceProvider / isAThing
(?:\b|_)(?:I|isA)(?=(?:[A-Z][a-z]{2,})+(?:[A-Z]|\b))

# hit-count: 6 file-count: 1
# in check-spelling@v0.0.22+, printf markers aren't automatically consumed
# printf markers
(?<!\\)\\[nrt](?=[a-z]{2,})

# hit-count: 6 file-count: 1
# alternate printf markers if you run into latex and friends
(?<!\\)\\[nrt](?=[a-z]{2,})(?=.*['"`])

# hit-count: 3 file-count: 3
# Google APIs
\b[-a-zA-Z0-9.]*\.googleapis\.com(?:\S*|)

# Jupyter Notebook ID
"id":\s*"[a-zA-Z0-9_-]{4,36}"

# Embedded Images
[A-Za-z0-9+\/=-]{10,}

# While you could try to match `http://` and `https://` by using `s?` in `https?://`, sometimes there
# YouTube url
\b(?:(?:www\.|)youtube\.com|youtu.be)/(?:channel/|embed/|user/|playlist\?list=|watch\?v=|v/|)[-a-zA-Z0-9?&=_%]*
# YouTube music
\bmusic\.youtube\.com/youtubei/v1/browse(?:[?&]\w+=[-a-zA-Z0-9?&=_]*)
# YouTube tag
<\s*youtube\s+id=['"][-a-zA-Z0-9?_]*['"]
# YouTube image
\bimg\.youtube\.com/vi/[-a-zA-Z0-9?&=_]*
# Google Accounts
\baccounts.google.com/[-_/?=.:;+%&0-9a-zA-Z]*
# Google Analytics
\bgoogle-analytics\.com/collect.[-0-9a-zA-Z?%=&_.~]*
# Google APIs
\bgoogleapis\.(?:com|dev)/[a-z]+/(?:v\d+/|)[a-z]+/[-@:./?=\w+|&]+
# Google Storage
\b[-a-zA-Z0-9.]*\bstorage\d*\.googleapis\.com(?:/\S*|)
# Google Calendar
\bcalendar\.google\.com/calendar(?:/u/\d+|)/embed\?src=[@./?=\w&%]+
\w+\@group\.calendar\.google\.com\b
# Google DataStudio
\bdatastudio\.google\.com/(?:(?:c/|)u/\d+/|)(?:embed/|)(?:open|reporting|datasources|s)/[-0-9a-zA-Z]+(?:/page/[-0-9a-zA-Z]+|)
# The leading `/` here is as opposed to the `\b` above
# ... a short way to match `https://` or `http://` since most urls have one of those prefixes
# Google Docs
/docs\.google\.com/[a-z]+/(?:ccc\?key=\w+|(?:u/\d+|d/(?:e/|)[0-9a-zA-Z_-]+/)?(?:edit\?[-\w=#.]*|/\?[\w=&]*|))
# Google Drive
\bdrive\.google\.com/(?:file/d/|open)[-0-9a-zA-Z_?=]*
# Google Groups
\bgroups\.google\.com(?:/[a-z]+/(?:#!|)[^/\s"]+)*
# Google Maps
\bmaps\.google\.com/maps\?[\w&;=]*
# Google themes
themes\.googleusercontent\.com/static/fonts/[^/\s"]+/v\d+/[^.]+.
# Google CDN
\bclients2\.google(?:usercontent|)\.com[-0-9a-zA-Z/.]*
# Goo.gl
/goo\.gl/[a-zA-Z0-9]+
# Google Chrome Store
\bchrome\.google\.com/webstore/detail/[-\w]*(?:/\w*|)
# Google Books
\bgoogle\.(?:\w{2,4})/books(?:/\w+)*\?[-\w\d=&#.]*
# Google Fonts
\bfonts\.(?:googleapis|gstatic)\.com/[-/?=:;+&0-9a-zA-Z]*
# Google Forms
\bforms\.gle/\w+
# Google Scholar
\bscholar\.google\.com/citations\?user=[A-Za-z0-9_]+
# Google Colab Research Drive
\bcolab\.research\.google\.com/drive/[-0-9a-zA-Z_?=]*

# GitHub SHAs (api)
\bapi.github\.com/repos(?:/[^/\s"]+){3}/[0-9a-f]+\b
# GitHub SHAs (markdown)
(?:\[`?[0-9a-f]+`?\]\(https:/|)/(?:www\.|)github\.com(?:/[^/\s"]+){2,}(?:/[^/\s")]+)(?:[0-9a-f]+(?:[-0-9a-zA-Z/#.]*|)\b|)
# GitHub SHAs
\bgithub\.com(?:/[^/\s"]+){2}[@#][0-9a-f]+\b
# GitHub SHA refs
\[([0-9a-f]+)\]\(https://(?:www\.|)github.com/[-\w]+/[-\w]+/commit/\g{-1}[0-9a-f]*
# GitHub wiki
\bgithub\.com/(?:[^/]+/){2}wiki/(?:(?:[^/]+/|)_history|[^/]+(?:/_compare|)/[0-9a-f.]{40,})\b
# githubusercontent
/[-a-z0-9]+\.githubusercontent\.com/[-a-zA-Z0-9?&=_\/.]*
# githubassets
\bgithubassets.com/[0-9a-f]+(?:[-/\w.]+)
# gist github
\bgist\.github\.com/[^/\s"]+/[0-9a-f]+
# git.io
\bgit\.io/[0-9a-zA-Z]+
# GitHub JSON
"node_id": "[-a-zA-Z=;:/0-9+_]*"
# Contributor
\[[^\]]+\]\(https://github\.com/[^/\s"]+/?\)
# GHSA
GHSA(?:-[0-9a-z]{4}){3}

# Twitter/X links
\b(twitter|x)\.com[\w/]+\b

# Hugging Face links
\bhuggingface.co[\w\./-]+\b

# GitHub actions
\buses:\s+[-\w.]+/[-\w./]+@[-\w.]+

# GitLab commit
\bgitlab\.[^/\s"]*/\S+/\S+/commit/[0-9a-f]{7,16}#[0-9a-f]{40}\b
# GitLab merge requests
\bgitlab\.[^/\s"]*/\S+/\S+/-/merge_requests/\d+/diffs#[0-9a-f]{40}\b
# GitLab uploads
\bgitlab\.[^/\s"]*/uploads/[-a-zA-Z=;:/0-9+]*
# GitLab commits
\bgitlab\.[^/\s"]*/(?:[^/\s"]+/){2}commits?/[0-9a-f]+\b

# Truncated outputs in Jupyter Notebooks
[A-Za-z]+\.\.\.

# mailto urls
mailto:[-a-zA-Z=;:/?%&0-9+@._]{3,}
