new-words

annotate new-words.sh @ 34:3827cce83602

find changesets by author, revision, files, or words in the commit message
smallfix
author igor@chub.in
date Tue Dec 21 15:53:22 2010 +0200 (2010-12-21)
parents 720a701b2ba9
children f95804355b0f
rev   line source
igor@26 1 #!/bin/bash
igor@20 2
igor@20 3 show_usage()
igor@20 4 {
igor@20 5 cat <<HELP > /dev/stderr
igor@0 6
igor@16 7 USAGE:
igor@16 8
igor@16 9 new-words [ -l lang ] [ -s ] [ ARG ]
igor@16 10
igor@16 11 SWITCHES:
igor@16 12
igor@20 13 -h print this screen
igor@20 14 -k put higher words that are similar to the known words (only for English)
igor@16 15 -l lang override language settings
igor@21 16 -n non-interactive mode (don't run vi)
igor@33 17 -N turn off known words filtering
igor@27 18 -a don't add marks (and don't save marks added by user)
igor@25 19 -p pages work with specified pages only (pages = start-stop/total )
igor@20 20 -s show the text statistics (percentage of known words and so on) and exit
igor@33 21 -S show your vocabulary statistics (number of words and word groups)
igor@27 22 -t tag tag known words with tag
igor@27 23 -T show list of active tags
igor@27 24 -m tag merge the words tagged with "tag" into the main vocabulary
igor@27 25 -M merge the words tagged with any tag into the main vocabulary
igor@27 26 -r tag remove subvocabulary for the "tag"
igor@26 27 -2 -3 find 2 and 3 words' sequences
igor@16 28
igor@20 29 The language of the text can be specified also
igor@20 30 by name of the program new-words (correspondent link must be created before).
igor@20 31 For example, these calls are equivalent:
igor@0 32
igor@20 33 de-words URL
igor@0 34 new-words -l de URL
igor@0 35
igor@20 36 HELP
igor@20 37 }
igor@0 38
igor@20 39 if [ "$1" = "-h" ]
igor@20 40 then
igor@20 41 show_usage
igor@20 42 exit 0
igor@20 43 fi
igor@0 44
igor@0 45 WORK_DIR=~/.new-words/
igor@28 46 TEMP1=`mktemp /tmp/new-words-temp1.XXXXXXXXXX`
igor@28 47 TEMP2=`mktemp /tmp/new-words-temp2.XXXXXXXXXX`
igor@28 48 export ORIGINAL_TEXT=`mktemp /tmp/new-words-orig.XXXXXXXXXX`
igor@0 49 editor=${EDITOR:-vim}
igor@0 50
igor@24 51 # language detection
igor@21 52
igor@2 53 LANGUAGE=en
igor@2 54 my_name="`echo $0 | sed s@.*/@@ | sed s/-.*// `"
igor@22 55 for arg
igor@22 56 do
igor@22 57 if echo "$arg" | grep -q http://...wikipedia.org/wiki/
igor@22 58 then
igor@22 59 LANGUAGE="`echo $arg | sed s@http://@@ | sed s@.wikipedia.*@@`"
igor@22 60 fi
igor@22 61 done
igor@2 62 [ "${my_name}" = "new" ] || LANGUAGE="$my_name"
igor@24 63
igor@24 64 #----------------------------------------------------
igor@24 65 # command line options processing
igor@24 66
igor@24 67 STAT_ONLY=NO
igor@24 68 NEED_TO_USE_VOCABULARY_WHEN_SORT=NO
igor@24 69 DONT_ADD_MARKS=NO
igor@24 70 NON_INTERACTIVE_MODE=NO
igor@25 71 PART_TO_PROCESS=''
igor@26 72 GROUP_WORDS_BY_THREE=NO
igor@26 73 GROUP_WORDS_BY_TWO=NO
igor@27 74 TAG_NAME=''
igor@27 75 MERGE_THIS_TAGS=''
igor@27 76 TAGS_LIST_ONLY=NO
igor@27 77 MERGE_TAGGED_WORDS=NO
igor@27 78 MERGE_ALL_TAGGED=NO
igor@27 79 DONT_ADD_MARKLINES=NO
igor@33 80 FILTER_WORDS=YES
igor@33 81 SHOW_VOC_STAT=NO
igor@33 82 while getopts l:sSkanNp:t:Tm:Mr:23 opt
igor@24 83 do
igor@24 84 case "$opt" in
igor@24 85 s) STAT_ONLY=YES;;
igor@33 86 S) SHOW_VOC_STAT=YES;;
igor@24 87 k) NEED_TO_USE_VOCABULARY_WHEN_SORT=YES;;
igor@24 88 l) LANGUAGE="$OPTARG";;
igor@27 89 a) DONT_ADD_MARKS=YES;;
igor@24 90 n) NON_INTERACTIVE_MODE=YES;;
igor@33 91 N) FILTER_WORDS=NO;;
igor@25 92 p) PART_TO_PROCESS="$OPTARG";;
igor@27 93 t) TAG_NAME="$OPTARG";;
igor@27 94 T) TAGS_LIST_ONLY="YES";;
igor@27 95 m) DONT_ADD_MARKLINES="YES"; MERGE_TAGGED_WORDS="YES"; MERGE_THIS_TAGS="$TAG_NAME $OPTARG";;
igor@27 96 M) DONT_ADD_MARKLINES="YES"; MERGE_ALL_TAGGED="YES";;
igor@27 97 r) REMOVE_TAG="YES"; TAG_NAME="$TAG_NAME $OPTARG";;
igor@26 98 2) GROUP_WORDS_BY_TWO=YES;;
igor@26 99 3) GROUP_WORDS_BY_THREE=YES;;
igor@24 100 \?) # unknown flag
igor@24 101 show_usage
igor@24 102 exit 1;;
igor@24 103 esac
igor@24 104 done
igor@24 105 shift `expr $OPTIND - 1`
igor@24 106
igor@0 107 if [ "$1" = "-l" ]
igor@0 108 then
igor@0 109 LANGUAGE="$2"
igor@0 110 shift 2
igor@0 111 fi
igor@24 112
igor@2 113 VOCABULARY=${LANGUAGE}.txt
igor@2 114 NOTES_FILE=notes-${LANGUAGE}.txt
igor@0 115
igor@33 116 if [ "${SHOW_VOC_STAT}" = "YES" ]
igor@33 117 then
igor@33 118 $0 -l "${LANGUAGE}" -N -n ${WORK_DIR}/${VOCABULARY} | head -1 | awk '{print $5}' | tr -d "<>"
igor@33 119 exit 0
igor@33 120 fi
igor@33 121
igor@21 122 #----------------------------------------------------
igor@21 123
igor@0 124 get_words()
igor@0 125 {
igor@33 126 export FILTER_WORDS
igor@1 127 tr ' ' '\n' | sed 's/--/ /g' \
igor@3 128 | sed "s/'/__APOSTROPHE__/g" \
igor@31 129 | perl -MEncode -Mutf8 -n -e '$_ = decode( "utf8", $_);y/*\r,.:#@()+=—<>$;"?!|·[]^%&/ /; binmode STDOUT, ":utf8"; print if /^[[:alpha:] '"'"'_-]*$/'\
igor@26 130 | sed "s/__APOSTROPHE__/'/g" \
igor@3 131 | tr ' ' '\n' \
igor@3 132 | tee "$1" \
igor@3 133 | grep_v_english_perl \
igor@0 134 | sort | uniq -c | awk '{if ($2!="") print;}' | sort -rn
igor@0 135 }
igor@0 136
igor@3 137 add_stat()
igor@3 138 {
igor@27 139 if [ "$DONT_ADD_MARKLINES" = "YES" ]
igor@27 140 then
igor@27 141 cat
igor@27 142 return
igor@27 143 fi
igor@3 144 before="$1"
igor@3 145 after=${before}2
igor@3 146 cat > "$after"
igor@3 147 total="`wc -w $1 | awk '{print $1}'`"
igor@3 148 total_unknown="`cat $after|awk '{s=s+$1}END{print s}'`"
igor@3 149 total_known="`echo $total-$total_unknown|bc`"
igor@3 150 percentage="`echo '100*('$total-$total_unknown')'/$total | bc -l | sed 's/\\.\(.\).*/.\1/'`"
igor@18 151 #sentences="`cat $after | perl -e 'local $/; $_=<>; s@http://[a-zA-Z&_.:/0-9%?=,\#+()\[\]~-]*@@g; s@\n@@g; s@(Mr|Mrs)\.@\1POINT@g; @sentences=split /\\./;print $#sentences;'`"
igor@18 152 sentences="`cat $ORIGINAL_TEXT | perl -e 'local $/; $_=<>; s/[^.]//msg; print length($_);'`"
igor@18 153
igor@18 154
igor@16 155 if [ "$STAT_ONLY" = "YES" ]
igor@16 156 then
igor@18 157 echo "LANG KNOWN% UNKNOWN% KNOWN TOTAL WPS UWPS*10"
igor@18 158 echo "$LANGUAGE $percentage `echo \(100-$percentage\) | bc -l` $total_known $total `echo $total/$sentences|bc` `echo 10*$total_unknown/$sentences|bc` "
igor@17 159 rm $after
igor@16 160 return 0
igor@16 161 else
igor@34 162 groups="`echo $(grep '# groups' $after | awk '{print $3}')`"
igor@34 163 words="`echo $(grep -v '^#' $after | wc -l)`"
igor@33 164 echo "# $LANGUAGE, $percentage, <$total_known/$total>, <$groups/$words>"
igor@16 165 fi
igor@16 166
igor@14 167 PERL_SCRIPT_TEMP_NAME=`mktemp /tmp/perl-grep-v-english-XXXXXXXX`
igor@14 168 cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME
igor@14 169 my $total=shift(@ARGV);
igor@14 170 my $total_known=shift(@ARGV);
igor@14 171 my $s=0;
igor@16 172 my $mark_line=int($total_known*100/$total/5)*5;
igor@19 173 if ($mark_line>=90) {
igor@19 174 $mark_line=int($total_known*100/$total)+1;
igor@19 175 } else { $mark_line +=5; };
igor@14 176 while(<>)
igor@14 177 {
igor@33 178 next if /^#\s*groups\s*/;
igor@14 179 print;
igor@14 180 /^\s*([0-9]*)\s*/;
igor@14 181 $s+=$1;
igor@19 182 if (($total_known+$s)*100/$total>=$mark_line) {
igor@14 183 print "# $mark_line\n";
igor@14 184 if ($mark_line>=90) { $mark_line+=1; } else { $mark_line +=5; };
igor@14 185 }
igor@14 186 }
igor@14 187 PERL_SCRIPT
igor@14 188 perl $PERL_SCRIPT_TEMP_NAME "$total" "$total_known" "$after"
igor@14 189 rm $PERL_SCRIPT_TEMP_NAME
igor@3 190 rm $after
igor@3 191 }
igor@3 192
igor@26 193 two_and_three_words()
igor@26 194 {
igor@26 195 if [ $GROUP_WORDS_BY_THREE = NO -a $GROUP_WORDS_BY_TWO = NO ]
igor@26 196 then
igor@26 197 cat
igor@26 198 else
igor@26 199 cat
igor@26 200
igor@26 201 export GROUP_WORDS_BY_THREE
igor@26 202 export GROUP_WORDS_BY_TWO
igor@26 203 PERL_SCRIPT_TEMP_NAME=`mktemp /tmp/perl-two-and-three-XXXXXXXX`
igor@26 204 cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME
igor@26 205 #!/usr/bin/perl
igor@26 206 local $/;
igor@26 207 $words=<>;
igor@26 208 $words=~ s@[!?;,:#1-9".]@ @g;
igor@26 209 $words =~ s@\s+@ @g;
igor@26 210 @words = split /\s+/, $words;
igor@26 211 for ($i=0; $i<$#words-3;$i++) {
igor@26 212 my ($a,$b,$c)= ($words[$i],$words[$i+1],$words[$i+2]);
igor@26 213 if ($ENV{GROUP_WORDS_BY_THREE} eq "YES" and ($a && $b && $c)) {
igor@26 214 print "${a}_${b}_${c}\n";
igor@26 215 };
igor@26 216 if ($ENV{GROUP_WORDS_BY_TWO} eq "YES" and ($a && $b)) {
igor@26 217 print "${a}_${b}\n";
igor@26 218 };
igor@26 219 }
igor@26 220 PERL_SCRIPT
igor@26 221 perl $PERL_SCRIPT_TEMP_NAME "$ORIGINAL_TEXT"
igor@26 222 rm $PERL_SCRIPT_TEMP_NAME
igor@26 223 fi
igor@26 224 }
igor@26 225
igor@0 226 grep_v_english()
igor@0 227 {
igor@0 228 [ -e "$VOCABULARY" ] || touch "$VOCABULARY"
igor@0 229 eval $(cat $VOCABULARY | tr -d "'" | xargs -n10 echo | tr ' ' '|' | sed 's/^/egrep -xv "RRRRRRR|/' | sed 's/$/"/' | tr '\n' '|')cat
igor@0 230 }
igor@0 231
igor@0 232 grep_v_english_perl()
igor@0 233 {
igor@0 234 PERL_SCRIPT_TEMP_NAME=`mktemp /tmp/perl-grep-v-english-XXXXXXXX`
igor@0 235 cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME
igor@33 236 if ($ENV{FILTER_WORDS} eq "NO") {
igor@33 237 while(<>) { print; }
igor@33 238 exit(0);
igor@33 239 }
igor@27 240 $voc_files=$ENV{VOC_FILES};
igor@27 241 $voc_files=~s@^ @@;
igor@27 242 for $voc_file (split /\s+/,$voc_files) {
igor@27 243 if (open(VOC, $voc_file)) {
igor@27 244 while (<VOC>){
igor@27 245 chomp;
igor@27 246 #s/'//g;
igor@27 247 $voc{$_}="1";
igor@27 248 }
igor@27 249 }
igor@0 250 }
igor@0 251 while(<>) {
igor@0 252 chomp;
igor@0 253 if (not defined($voc{$_})) { print "$_\n"; }
igor@0 254 }
igor@0 255 PERL_SCRIPT
igor@0 256 [ -e "$VOCABULARY" ] || touch "$VOCABULARY"
igor@27 257 export VOCABULARY VOC_FILES
igor@27 258 VOC_FILES=$VOCABULARY
igor@27 259 for i in $TAG_NAME
igor@27 260 do
igor@27 261 VOC_FILES="${VOC_FILES} `tag_file_name $i`"
igor@27 262 done
igor@0 263 perl $PERL_SCRIPT_TEMP_NAME
igor@0 264 rm $PERL_SCRIPT_TEMP_NAME
igor@0 265 }
igor@0 266
igor@11 267 group_words()
igor@11 268 {
igor@22 269 #if [ "$LANGUAGE" != "en" ]
igor@22 270 #then
igor@22 271 # cat
igor@22 272 # return
igor@22 273 #fi
igor@11 274 PERL_SCRIPT_TEMP_NAME=`mktemp /tmp/perl-group-words-XXXXXXXX`
igor@11 275 cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME
igor@11 276 #!/usr/bin/perl
igor@11 277
igor@29 278 use Encode;
igor@29 279 use utf8;
igor@32 280 use Lingua::Stem::Snowball qw(stem);
igor@29 281
igor@12 282 eval {
igor@12 283 # http://stackoverflow.com/questions/251694/how-can-i-check-if-i-have-a-perl-module-before-using-it
igor@12 284 require String::Similarity;
igor@12 285 String::Similarity->import();
igor@12 286 };
igor@12 287 unless($@)
igor@12 288 {
igor@12 289 our $HAVE_String_Similarity=1;
igor@12 290 }
igor@12 291
igor@32 292
igor@32 293 sub load_notes_dict()
igor@32 294 {
igor@32 295 my %dict;
igor@32 296 if (open(NOTES, $ENV{NOTES_FILE})) {
igor@32 297 while(<NOTES>) {
igor@32 298 $_ = decode( "utf8", $_);
igor@32 299 chomp;
igor@32 300 s/^\s+//;
igor@32 301 my ($a,$b)=split /\s+/,$_,2;
igor@32 302 $dict{$a}=$b;
igor@32 303 }
igor@32 304 }
igor@32 305 return %dict;
igor@32 306 }
igor@32 307
igor@12 308 sub similar($$){
igor@12 309 my $a=shift;
igor@12 310 my $b=shift;
igor@12 311 if ($HAVE_String_Similarity) {
igor@12 312 return $Similarity{"$a $b"};
igor@12 313 }
igor@12 314 else {
igor@12 315 return 0;
igor@12 316 }
igor@12 317 }
igor@12 318
igor@32 319
igor@32 320 sub normalize_without_linked($)
igor@11 321 {
igor@22 322 if ( $ENV{LANGUAGE} eq "en" ) { return normalize_english(shift); }
igor@22 323 elsif ( $ENV{LANGUAGE} eq "de" ) { return normalize_german(shift); }
igor@29 324 elsif ( $ENV{LANGUAGE} eq "uk" ) { return normalize_ukrainian(shift); }
igor@31 325 elsif ( $ENV{LANGUAGE} eq "io" ) { return normalize_esperanto(shift); }
igor@22 326 else { return shift ; }
igor@22 327 }
igor@22 328
igor@32 329 sub normalize_with_linked($)
igor@32 330 {
igor@32 331 my $word = normalize_without_linked(shift);
igor@32 332 #return $word;
igor@32 333 if ($linked_words{$word}) {
igor@32 334 return $linked_words{$word};
igor@32 335 }
igor@32 336 else {
igor@32 337 return $word;
igor@32 338 }
igor@32 339 }
igor@32 340
igor@32 341 sub normalize($)
igor@32 342 {
igor@32 <