|
ajoute scriptt renomme
|
1 |
#!/bin/bash |
| 2 | ||
| 3 |
# script pour renommer des personnages ou lieux dans un document Manuskript |
|
| 4 |
# ./renomme <document manuskript> <ancien nom> <nouveau nom> [ins=poids] [del=poids] [rep=poids] |
|
| 5 |
# |
|
| 6 |
# Le document d'origine est sauvegardé (voir variable $backup) |
|
| 7 |
# |
|
| 8 |
# Si <nouveau nom> est "check", ou "prox" ou "leven", une étude de proximité est alors effectuée |
|
| 9 |
# sur l'algorithme Levenshtein. Les paramètres de poids sont des nombres entiers, et permettent |
|
| 10 |
# de pondérer l'ajout (ins=), la suppression (del=) et le remplacement (rep=) de caractère. |
|
| 11 |
# par défaut chacun des trois paramètres est égal à 1. |
|
| 12 |
# Il n'y a pas d'ordre obligatoire pour le paramétrage, et si un paramétrage est effectué plusieurs fois |
|
| 13 |
# c'est le plus à gauche qui prend la priorité. Il n'y a pas de backup effectué pour l'opération de vérification de la proximité |
|
| 14 |
# crédit algo: https://en.wikibooks.org/wiki/Algorithm_Implementation/Strings/Levenshtein_distance |
|
| 15 | ||
| 16 |
set -e |
|
| 17 | ||
| 18 |
manuscrit=${1:?}
|
|
| 19 |
ancien=${2:?}
|
|
| 20 |
nouveau=${3:?}
|
|
| 21 | ||
| 22 |
if test "${manuscrit:0:1}" != '/'; then
|
|
| 23 |
manuscrit="$PWD/$manuscrit" |
|
| 24 |
fi |
|
| 25 |
test -r "$manuscrit" || exit 1 |
|
| 26 |
test $(file --brief --mime-type --dereference "$manuscrit") == 'application/zip' || exit 2 |
|
| 27 | ||
| 28 |
backup="${manuscrit/%.msk} (avant renommage de «${ancien}» en «${nouveau}»).msk"
|
|
| 29 | ||
| 30 |
function trap_exit () {
|
|
| 31 |
rm -fr $temp |
|
| 32 |
cd - > /dev/null |
|
| 33 |
} |
|
| 34 | ||
| 35 |
function determinant () {
|
|
| 36 |
eval "local nom=\"\$$1\"" |
|
| 37 |
if [[ ${nom:0:1} == @(A|E|H|I|O|U) \
|
|
| 38 |
|| ${nom:0:1} == @(Â|Ê|H|Î|Ô|Û) \
|
|
| 39 |
|| ${nom:0:1} == @(Ä|Ë|H|Ï|Ö|Ü) \
|
|
| 40 |
|| ${nom:0:1} == @(À|È|H|Ì|Ò|Ù) \
|
|
| 41 |
|| ${nom:0:1} == @(Á|É|H|Í|Ó|Ú) \
|
|
| 42 |
]]; then |
|
| 43 |
eval "determinant_$1=\"d'\"" |
|
| 44 |
eval "determinant_$1_formatted=\"d-\"" |
|
| 45 |
else |
|
| 46 |
eval "determinant_$1=\"de \"" |
|
| 47 |
eval "determinant_$1_formatted=\"de-\"" |
|
| 48 |
fi |
|
| 49 |
} |
|
| 50 | ||
| 51 |
function format () {
|
|
| 52 |
eval "$1_formatted=\$(tr --complement --squeeze-repeats 'A-Za-z-_\n' - <<< \"\${$1// /_}\")"
|
|
| 53 |
} |
|
| 54 | ||
| 55 |
function renomme_fichiers () {
|
|
| 56 |
for char in $(find characters -type f -regex "characters/[0-9]+-$1.txt"); do |
|
| 57 |
local new_char=$(sed "s/$1/$2/" <<< $char) |
|
| 58 |
echo "character: $char -> $new_char" |
|
| 59 |
mv $char $new_char |
|
| 60 |
break |
|
| 61 |
done |
|
| 62 | ||
| 63 |
for chapter in $(find outline -type d -regex "outline/.*$1.*"); do |
|
| 64 |
local new_chapter=$(sed "s/$1/$2/g" <<< $chapter) |
|
| 65 |
echo "chapter: $chapter -> $new_chapter" |
|
| 66 |
mv $chapter $new_chapter |
|
| 67 |
done |
|
| 68 | ||
| 69 |
for part in $(find outline -type f -regex "outline/[^/]*$1[^/]*.md"); do |
|
| 70 |
local new_part=$(sed "s/$1/$2/g" <<< $part) |
|
| 71 |
echo "part: $part -> $new_part" |
|
| 72 |
mv $part $new_part |
|
| 73 |
done |
|
| 74 |
} |
|
| 75 | ||
| 76 |
trap trap_exit EXIT |
|
| 77 | ||
| 78 |
temp=$(mktemp --dry-run /dev/shm/XXXXXXXXX) |
|
| 79 |
mkdir $temp |
|
| 80 |
cd $temp |
|
| 81 | ||
| 82 |
if [[ $nouveau = @(check|prox|leven) ]]; then |
|
| 83 |
unzip -qq "$manuscrit" |
|
| 84 |
for param in $(seq 4 $#); do |
|
| 85 |
for dst in ins del rep; do |
|
| 86 |
eval "if test -n '\$$param' && [[ \"\$$param\" =~ $dst=[0-9]+ ]]; then cost_$dst=\${$param#*=}; fi"
|
|
| 87 |
done |
|
| 88 |
done |
|
| 89 |
echo paramètres d\'approximation |
|
| 90 |
echo "caractère manquant (del=): ${cost_del:-1}"
|
|
| 91 |
echo "caractère inséré (ins=): ${cost_ins:-1}"
|
|
| 92 |
echo "caractère remplacé (rep=): ${cost_rep:-1}"
|
|
| 93 |
for f in $(find . -type f); do |
|
| 94 |
let wc+=$(wc -w < $f) |
|
| 95 |
done |
|
| 96 |
awk -v ancien=$ancien -v wc=$wc -v cost_ins=${cost_ins:-1} -v cost_del=${cost_del:-1} -v cost_rep=${cost_rep:-1} '
|
|
| 97 |
BEGIN {
|
|
| 98 |
RS="[[:punct:]]" |
|
| 99 |
progress_mod = 10 |
|
| 100 |
actual_progress = 0 |
|
| 101 |
pct_progress = 0 |
|
| 102 |
progress = 0 |
|
| 103 |
found_words = 0 |
|
| 104 | ||
| 105 |
str1_len = length(ancien) |
|
| 106 |
for (i=1; i<=str1_len; i++) |
|
| 107 |
str1_substr[i]=substr(ancien, i, 1) |
|
| 108 |
} |
|
| 109 |
function levenshtein(str2) {
|
|
| 110 |
str2_len = length(str2) |
|
| 111 |
if(str2_len == 0) return str1_len * cost_del |
|
| 112 |
for(j = 1; j <= str2_len; j++) |
|
| 113 |
str2_substr[j]=substr(str2, j, 1) |
|
| 114 |
matrix[0, 0] = 0 |
|
| 115 |
for(i = 1; i <= str1_len; i++) {
|
|
| 116 |
matrix[i, 0] = i * cost_del |
|
| 117 |
for(j = 1; j <= str2_len; j++) {
|
|
| 118 |
matrix[0, j] = j * cost_ins |
|
| 119 |
x = matrix[i - 1, j] + cost_del |
|
| 120 |
y = matrix[i, j - 1] + cost_ins |
|
| 121 |
z = matrix[i - 1, j - 1] + (str1_substr[i] == str2_substr[j] ? 0 : cost_rep) |
|
| 122 |
x = x < y ? x : y |
|
| 123 |
matrix[i, j] = x < z ? x : z |
|
| 124 |
} |
|
| 125 |
} |
|
| 126 |
return matrix[str1_len, str2_len] |
|
| 127 |
} |
|
| 128 |
{
|
|
| 129 |
for (word=1; word<=NF; word++) {
|
|
| 130 |
progress++ |
|
| 131 |
lvstn = levenshtein(gensub("[[:punct:]]","","g",$word))
|
|
| 132 |
if (lvstn <= 3 && lvstn > 0) {
|
|
| 133 |
approx_possibles[$word]++ |
|
| 134 |
found_words++ |
|
| 135 |
} |
|
| 136 |
pct_progress=int(progress / wc * 100) |
|
| 137 |
if (actual_progress < pct_progress && pct_progress % progress_mod == 0) {
|
|
| 138 |
actual_progress = pct_progress |
|
| 139 |
printf("%i%\n", actual_progress)
|
|
| 140 |
} |
|
| 141 |
} |
|
| 142 |
} |
|
| 143 |
END {
|
|
| 144 |
if (found_words > 0) {
|
|
| 145 |
printf("mot%s proche%s de «%s» [occurences]\n", found_words > 0 ? "s" : "", found_words > 0 ? "s" : "", ancien)
|
|
| 146 |
for (i in approx_possibles) |
|
| 147 |
printf("- %s [%i]\n", i, approx_possibles[i])
|
|
| 148 |
} |
|
| 149 |
else {
|
|
| 150 |
print "aucun mot proche et différent de «" ancien "» trouvé" |
|
| 151 |
} |
|
| 152 |
} |
|
| 153 |
' $(find . -type f) |
|
| 154 |
exit |
|
| 155 |
fi |
|
| 156 | ||
| 157 |
mv --backup=numbered "$manuscrit" "$backup" |
|
| 158 | ||
| 159 |
unzip -qq "$backup" |
|
| 160 | ||
| 161 |
for version in ancien nouveau; do |
|
| 162 |
format $version |
|
| 163 |
determinant $version |
|
| 164 |
done |
|
| 165 | ||
| 166 |
declare -A remplacement |
|
| 167 |
remplacement=( |
|
| 168 |
[$ancien]="$nouveau" |
|
| 169 |
[$ancien_formatted]="$nouveau_formatted" |
|
| 170 |
[$determinant_ancien$ancien]="$determinant_nouveau$nouveau" |
|
| 171 |
[$determinant_ancien_formatted$ancien_formatted]="$determinant_nouveau_formatted$nouveau_formatted" |
|
| 172 |
) |
|
| 173 | ||
| 174 |
renomme_fichiers "$ancien_formatted" "$nouveau_formatted" |
|
| 175 |
renomme_fichiers "$determinant_ancien_formatted$ancien_formatted" "$determinant_nouveau_formatted$nouveau_formatted" |
|
| 176 | ||
| 177 |
egrep --word-regexp --only-matching --recursive --regexp="($determinant_ancien|$determinant_ancien_formatted)\?($ancien|$ancien_formatted)" . \ |
|
| 178 |
| awk -v name="$1" -F ':' ' |
|
| 179 |
{
|
|
| 180 |
if ($NF > 0) {
|
|
| 181 |
file[$1]++ |
|
| 182 |
nb++ |
|
| 183 |
} |
|
| 184 |
} |
|
| 185 |
END {
|
|
| 186 |
printf("remplacement de %i occurences pour %s dans %i fichiers\n", nb, name, asort(file))
|
|
| 187 |
}' |
|
| 188 | ||
| 189 |
for regexp in "${!remplacement[@]}"; do
|
|
| 190 |
egrep --word-regexp --files-with-matches --recursive --regexp="$regexp" . \ |
|
| 191 |
| xargs --no-run-if-empty sed --regexp-extended --in-place "s/(\W|^)$regexp(\W|$)/\1${remplacement[$regexp]}\2/g"
|
|
| 192 |
done |
|
| 193 | ||
| 194 |
zip --recurse-paths --no-dir-entries -qq "${manuscrit}" *
|