#!/usr/bin/env bash
###########################################################################
# check_spelling.sh
# ---------------------
# Date : December 2016
# Copyright : (C) 2016 by Denis Rouzaud
# Email : denis.rouzaud@gmail.com
###########################################################################
# #
# This program is free software; you can redistribute it and/or modify #
# it under the terms of the GNU General Public License as published by #
# the Free Software Foundation; either version 2 of the License, or #
# (at your option) any later version. #
# #
###########################################################################
# -r: deactivate interactive mode to fix errors
# optional argument: list of files to be checked
DIR=$(git rev-parse --show-toplevel)/scripts/spell_check
AGIGNORE=${DIR}/.agignore
# GNU prefix command for mac os support (gsed, gsplit)
GP=
if [[ "$OSTYPE" =~ darwin* ]]; then
GP=g
fi
# ARGUMENTS
INTERACTIVE=$( tty -s && echo YES || echo NO)
DEBUG=NO
OUTPUTLOG=""
while getopts ":rdl:" opt; do
case $opt in
r)
INTERACTIVE=NO
;;
d)
DEBUG=YES
;;
l)
OUTPUTLOG=$OPTARG
;;
\?)
echo "Invalid option: -$OPTARG" >&2
exit 1
;;
esac
done
shift $(expr $OPTIND - 1)
if [ ! $# -eq 0 ]; then
EXCLUDE=$(cat $AGIGNORE |${GP}sed -e 's/\s*#.*$//' -e '/^\s*$/d' | tr '\n' '|' |${GP}sed -e 's/|$//')
INPUTFILES=$(echo $@ | tr -s '[[:blank:]]' '\n' | egrep -iv "$EXCLUDE" | tr '\n' ' ' )
if [[ -z $INPUTFILES ]]; then
exit 0
fi
echo "Running spell check on files: $INPUTFILES"
else
INPUTFILES="."
fi
# regex to find escape string
SPELLOKRX='(#\s*spellok|<!--\s*#\s*spellok\s*-->)$'
# split into several files to avoid too long regexes
SPLIT=6
${GP}split --number=l/$SPLIT --numeric-suffixes --suffix-length=2 --additional-suffix=~ ${DIR}/spelling.dat spelling
# global replace variables (dictionary)
declare -A GLOBREP_ALLFILES=()
declare -A GLOBREP_CURRENTFILE=()
declare -A GLOBREP_IGNORE=()
ERRORFOUND=NO
for I in $(seq -f '%02g' 0 $(($SPLIT-1)) ) ; do
( [[ "$INTERACTIVE" =~ YES ]] || [[ "$TRAVIS" =~ true ]] ) && printf "Progress: %d/%d\n" $I $SPLIT
SPELLFILE=spelling$I~
# if correction contains an uppercase letter and is the same as the error character wise, this means that the error is searched as a full word and case sensitive (not incorporated in a bigger one)
CASEMATCH_FIXCASE=$(${GP}sed -rn '/^(\w+):\1(:\*)?$/Ip' $SPELLFILE | ${GP}sed -r 's/^(\w+):\1(:\*)?$/(\\b|_)\1(\\b|_)/I')
REMAINS=$( ${GP}sed -r '/^(\w+):\1(:\*)?$/Id' $SPELLFILE)
# for error or correction containing any non letter character (space, apostrophe) search is full word and case insensitive
IGNORECASE_FIXSPECIALCHAR=$(echo "$REMAINS" | perl -ne "print if /^(\w*[ '])*\w*:\w+(?(1)|[' ])/" | ${GP}sed -r 's/^(.*?):.*?(:\*)?$/(\\b|_)\1(\\b|_)/' )
REMAINS=$( echo "$REMAINS" | perl -ne "print if not /^(\w*[ '])*\w*:\w+(?(1)|[' ])/")
# This will try to look for misspelling within larger words.
# Condition is hard to explain in words.
# You can test it here: https://regex101.com/r/7kznVA/12
# adding :* in spelling.dat willextra words that should not be checked in longer words ca
# remove those in spelling.dat ending with :*
# following can be checked in longer words case insensitively
IGNORECASE_INWORD=$(echo "$REMAINS" | perl -ne 'print if /^(\w)(\w)(\w)\w*(\w)(\w)(\w):(?:(?!\2\3\w|\w\1\2).)\w*?(?:(?!\5\6\w|\w\4\5)\w\w\w)$/' | cut -d: -f1 )
REMAINS=$( echo "$REMAINS" | perl -ne 'print if not /^(\w)(\w)(\w)\w*(\w)(\w)(\w):(?:(?!\2\3\w|\w\1\2).)\w*?(?:(?!\5\6\w|\w\4\5)\w\w\w)$/' | cut -d: -f1 )
# Trying with the rest as whole words case insensitively
IGNORECASE_WHOLEWORD=$(echo "$REMAINS" | ${GP}sed -r 's/^/(\\b|_)/; s/$/(\\b|_)/' )
# or in camel case, case sensitively for word of at least 4 chars
MATCHCASE_INWORD=$(echo "$REMAINS" | ${GP}sed -r '/^.{,3}$/d' | ${GP}sed -r 's/^(\w)(.*)/(\\b|_)(\l\1\L\2_|\u\1\U\2_|\u\1\L\2\U[_A-Z0-9])|\L[_a-z0-9]\u\1\L\2(\\b|\U[_A-Z0-9])|\L[_a-z0-9]\u\1\U\2\L(\\b|[_a-z0-9])/' )
if [[ "$DEBUG" =~ YES ]]; then
echo "*** FIX CASE (case sensitive) ***"
echo "$CASEMATCH_FIXCASE"
echo "*** SPECIAL CHAR (case insensitive) ***"
echo "$IGNORECASE_FIXSPECIALCHAR"
echo "*** IN WORD (case insensitive) ***"
echo "$IGNORECASE_INWORD"
echo "*** WHOLE WORDS (case insensitive) ***"
echo "$IGNORECASE_WHOLEWORD"
echo "*** IN WORD CAMELCASE (case sensitive) **"
echo "$MATCHCASE_INWORD"
echo "*****"
fi
CASEMATCH_FIXCASE=$( echo "$CASEMATCH_FIXCASE" | ${GP}sed -r '/^\s*$/d' | tr '\n' '\|' | ${GP}sed -r 's/\|$//')
IGNORECASE_FIXSPECIALCHAR=$(echo "$IGNORECASE_FIXSPECIALCHAR" | ${GP}sed -r '/^\s*$/d' | tr '\n' '\|' | ${GP}sed -r 's/\|$//')
IGNORECASE_INWORD=$( echo "$IGNORECASE_INWORD" | ${GP}sed -r '/^\s*$/d' | tr '\n' '\|' | ${GP}sed -r 's/\|$//')
IGNORECASE_WHOLEWORD=$( echo "$IGNORECASE_WHOLEWORD" | ${GP}sed -r '/^\s*$/d' | tr '\n' '\|' | ${GP}sed -r 's/\|$//')
MATCHCASE_INWORD=$( echo "$MATCHCASE_INWORD" | ${GP}sed -r '/^\s*$/d' | tr '\n' '\|' | ${GP}sed -r 's/\|$//')
IGNORECASE=$(echo "(${IGNORECASE_FIXSPECIALCHAR}|${IGNORECASE_INWORD}|${IGNORECASE_WHOLEWORD})" |${GP}sed -r 's/\(\|/(/' |${GP}sed -r 's/\|\|/|/g' |${GP}sed -r 's/\|\)/)/')'(?!.*'"${SPELLOKRX}"')'
CASEMATCH=$(echo "(${CASEMATCH_FIXCASE}|${MATCHCASE_INWORD})" |${GP}sed -r 's/\(\|/(/' |${GP}sed -r 's/\|\|/|/g' |${GP}sed -r 's/\|\)/)/')'(?!.*'"${SPELLOKRX}"')'
FILE=$INPUTFILES # init with input files (if ag is run with single file, file path is now in output)
while read -u 3 -r LINE; do
echo "$LINE"
ERRORFOUND=YES
NOCOLOR=$(echo "$LINE" | ${GP}sed -r 's/\x1B\[([0-9]{1,2}(;[0-9]{1,2})?)?[m|K]//g')
if [[ "$NOCOLOR" =~ ^[[:alnum:]][[:alnum:]\/\._-]+$ ]]; then
FILE=$NOCOLOR
GLOBREP_CURRENTFILE=()
fi
if [[ "$NOCOLOR" =~ ^[0-9]+: ]]; then
if [[ -z $FILE ]]; then
echo "*** error: no file"
exit 1
fi
NUMBER=$(echo "$NOCOLOR" | cut -d: -f1)
ERRORLINE=$(echo "$NOCOLOR" | cut -d: -f2)
ERROR=$(echo "$LINE" | ${GP}sed -r 's/^.*?\x1B\[30;43m(.*?)\x1B\[0m.*$/\1/')
ERRORNOCOLOR=$(echo "$ERROR" | ${GP}sed -r 's/\x1B\[([0-9]{1,2}(;[0-9]{1,2})?)?[m|K]//g')
if [[ "$ERRORNOCOLOR" =~ ^[[:digit:]]+: ]]; then
echo "*** error: could not find error in $LINE" >&2
else
# if the error is not in IGNORECASE_INWORD, then it matched previous and next character (needs to remove them)
# also make error small case and escape special chars: ( ) |
ERRORSMALLCASE=$(echo ${ERROR,,} |${GP}sed -r 's/\(/\\(/g' |${GP}sed -r 's/\)/\\)/g' |${GP}sed -r 's/\|/\\|/g')
if [[ ! "${ERRORSMALLCASE}" =~ $IGNORECASE_INWORD ]]; then
if [[ ! -z $(ag --nonumbers --case-sensitive "^${ERRORSMALLCASE:1:-1}${ERRORSMALLCASE: -1}?:" scripts/spell_check/spelling.dat) ]]; then
ERRORSMALLCASE=${ERRORSMALLCASE#?}
ERROR=${ERROR#?}
fi
if [[ ! -z $(ag --nonumbers --case-sensitive "^${ERRORSMALLCASE::-1}:" scripts/spell_check/spelling.dat) ]]; then
ERRORSMALLCASE=${ERRORSMALLCASE::-1}
ERROR=${ERROR::-1}
fi
fi
# get correction from spelling.dat
CORRECTION=$(ag --nonumbers --case-sensitive "^${ERRORSMALLCASE}:" ${DIR}/spelling.dat | cut -d: -f2)
if [[ -z $CORRECTION ]]; then
echo "could not find correction for $ERROR" >&2
else
# Match case
MATCHCASE="$ERROR:$CORRECTION"
CORRECTIONCASE=$(echo "$MATCHCASE" | ${GP}sed -r 's/([A-Z]+):(.*)/\1:\U\2/; s/([A-Z][a-z]+):([a-z])/\1:\U\2\L/' | cut -d: -f2)
if [[ ! -z $OUTPUTLOG ]]; then
echo "$FILE $NUMBER $ERROR $CORRECTIONCASE" >> $OUTPUTLOG
fi
if [[ "$INTERACTIVE" =~ YES ]]; then
# Skip global replace
if [[ ! -z ${GLOBREP_ALLFILES["$ERROR"]} ]]; then
echo -e "replace \x1B[33m$ERROR\x1B[0m by \x1B[33m$CORRECTIONCASE\x1B[0m in \x1B[33m$FILE\x1B[0m"
${GP}sed -i -r "/${SPELLOKRX}/! s/$ERROR/$CORRECTIONCASE/g" $FILE
continue
elif [[ ( ! -z ${GLOBREP_CURRENTFILE["$ERROR"]} ) || ( ! -z ${GLOBREP_IGNORE["$ERROR"]} ) ]]; then
echo "skipping occurrence"
continue
else
# escape string
SPELLOKSTR='//#spellok'
if [[ "$FILE" =~ \.(txt|html|htm|dox)$ ]]; then
SPELLOKSTR='<!--#spellok-->'
fi
if [[ "$FILE" =~ \.(h|cpp|sip)$ ]]; then
if [[ "$ERRORLINE" =~ ^\s*(\/*\|\/\/) ]]; then
SPELLOKSTR='#spellok'
fi
fi
if [[ "$FILE" =~ \.(py)$ ]]; then
SPELLOKSTR='#spellok'
fi
SPELLOKSTR_ESC=$(echo "$SPELLOKSTR" | ${GP}sed -r 's/\//\\\//g')
# Display menu
echo "***"
echo -e "Error found: \x1B[31m$ERROR\x1B[0m"
echo -e " r) \x1B[4mr\x1B[0meplace by \x1B[33m$CORRECTIONCASE\x1B[0m at line $NUMBER"
echo -e " f) replace all occurrences by \x1B[33m$CORRECTIONCASE\x1B[0m in current \x1B[4mf\x1B[0mile"
echo -e " a) replace all occurrences by \x1B[33m$CORRECTIONCASE\x1B[0m in \x1B[4ma\x1B[0mll files"
echo -e " p) a\x1B[4mp\x1B[0mpend \x1B[33m$SPELLOKSTR\x1B[0m at the end of the line $NUMBER to avoid spell check on this line"
echo -e " t) \x1B[4mt\x1B[0mype your own correction"
echo -e " c) skip and \x1B[4mc\x1B[0montinue"
echo -e " o) skip all \x1B[4mo\x1B[0mccurences and continue"
echo -e " e) \x1B[4me\x1B[0mxit"
while read -n 1 n; do
echo ""
case $n in
r)
echo -e "replacing \x1B[33m$ERROR\x1B[0m by \x1B[33m$CORRECTIONCASE\x1B[0m in \x1B[33m$FILE\x1B[0m at line \x1B[33m$NUMBER\x1B[0m"
${GP}sed -i "${NUMBER}s/$ERROR/$CORRECTIONCASE/g" $FILE
break
;;
f)
GLOBREP_CURRENTFILE+=(["$ERROR"]=1)
echo -e "replacing \x1B[33m$ERROR\x1B[0m by \x1B[33m$CORRECTIONCASE\x1B[0m in \x1B[33m$FILE\x1B[0m"
${GP}sed -i -r "/${SPELLOKRX}/! s/$ERROR/$CORRECTIONCASE/g" $FILE
break
;;
a)
GLOBREP_CURRENTFILE+=(["$ERROR"]=1)
GLOBREP_ALLFILES+=(["$ERROR"]=1)
echo -e "replace \x1B[33m$ERROR\x1B[0m by \x1B[33m$CORRECTIONCASE\x1B[0m in \x1B[33m$FILE\x1B[0m"
${GP}sed -i -r "/${SPELLOKRX}/! s/$ERROR/$CORRECTIONCASE/g" $FILE
break
;;
p)
echo -e "appending \x1B[33m$SPELLOKSTR\x1B[0m to \x1B[33m$FILE\x1B[0m at line \x1B[33m$NUMBER\x1B[0m"
${GP}sed -i "${NUMBER}s/\$/ $SPELLOKSTR_ESC/" $FILE
break
;;
t)
echo "Enter the correction: "
read CORRECTION
MATCHCASE="$ERROR:$CORRECTION"
CORRECTIONCASE=$(echo "$MATCHCASE" | ${GP}sed -r 's/([A-Z]+):(.*)/\1:\U\2/; s/([A-Z][a-z]+):([a-z])/\1:\U\2\L/' | cut -d: -f2)
echo -e "replacing \x1B[33m$ERROR\x1B[0m by \x1B[33m$CORRECTIONCASE\x1B[0m in \x1B[33m$FILE\x1B[0m at line \x1B[33m$NUMBER\x1B[0m"
${GP}sed -i "${NUMBER}s/$ERROR/$CORRECTIONCASE/g" $FILE
break
;;
c)
break
;;
o)
GLOBREP_IGNORE+=(["$ERROR"]=1)
break
;;
e)
exit 1
;;
*) invalid option;;
esac
done
fi
fi
fi
fi
if [[ "$NOCOLOR" =~ ^\s*$ ]]; then
FILE=""
fi
fi
done 3< <(
unbuffer ag --all-text --nopager --color-match "30;43" --numbers --nomultiline --ignore-case -p $AGIGNORE "${IGNORECASE}" $INPUTFILES
unbuffer ag --all-text --nopager --color-match "30;43" --numbers --nomultiline --case-sensitive -p $AGIGNORE "${CASEMATCH}" $INPUTFILES
)
rm -f $SPELLFILE
done
if [[ "$ERRORFOUND" =~ YES ]]; then
echo -e "\x1B[1msome errors have been found.\x1B[0m" >&2
exit 1
else
exit 0
fi