new-words

annotate new-words.sh @ 36:f95804355b0f

find changesets by author, revision, files, or words in the commit message
compressed wordlist support
author Igor Chubin <igor@chub.in>
date Sat Jan 01 19:47:39 2011 +0100 (2011-01-01)
parents 3827cce83602
children 4e931db74618
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@36 14 -c show compressed wordlist: one word per group
igor@20 15 -k put higher words that are similar to the known words (only for English)
igor@16 16 -l lang override language settings
igor@21 17 -n non-interactive mode (don't run vi)
igor@33 18 -N turn off known words filtering
igor@27 19 -a don't add marks (and don't save marks added by user)
igor@25 20 -p pages work with specified pages only (pages = start-stop/total )
igor@20 21 -s show the text statistics (percentage of known words and so on) and exit
igor@33 22 -S show your vocabulary statistics (number of words and word groups)
igor@27 23 -t tag tag known words with tag
igor@27 24 -T show list of active tags
igor@27 25 -m tag merge the words tagged with "tag" into the main vocabulary
igor@27 26 -M merge the words tagged with any tag into the main vocabulary
igor@27 27 -r tag remove subvocabulary for the "tag"
igor@26 28 -2 -3 find 2 and 3 words' sequences
igor@16 29
igor@20 30 The language of the text can be specified also
igor@20 31 by name of the program new-words (correspondent link must be created before).
igor@20 32 For example, these calls are equivalent:
igor@0 33
igor@20 34 de-words URL
igor@0 35 new-words -l de URL
igor@0 36
igor@20 37 HELP
igor@20 38 }
igor@0 39
igor@20 40 if [ "$1" = "-h" ]
igor@20 41 then
igor@20 42 show_usage
igor@20 43 exit 0
igor@20 44 fi
igor@0 45
igor@0 46 WORK_DIR=~/.new-words/
igor@28 47 TEMP1=`mktemp /tmp/new-words-temp1.XXXXXXXXXX`
igor@28 48 TEMP2=`mktemp /tmp/new-words-temp2.XXXXXXXXXX`
igor@28 49 export ORIGINAL_TEXT=`mktemp /tmp/new-words-orig.XXXXXXXXXX`
igor@0 50 editor=${EDITOR:-vim}
igor@0 51
igor@24 52 # language detection
igor@21 53
igor@2 54 LANGUAGE=en
igor@2 55 my_name="`echo $0 | sed s@.*/@@ | sed s/-.*// `"
igor@22 56 for arg
igor@22 57 do
igor@22 58 if echo "$arg" | grep -q http://...wikipedia.org/wiki/
igor@22 59 then
igor@22 60 LANGUAGE="`echo $arg | sed s@http://@@ | sed s@.wikipedia.*@@`"
igor@22 61 fi
igor@22 62 done
igor@2 63 [ "${my_name}" = "new" ] || LANGUAGE="$my_name"
igor@24 64
igor@24 65 #----------------------------------------------------
igor@24 66 # command line options processing
igor@24 67
igor@24 68 STAT_ONLY=NO
igor@24 69 NEED_TO_USE_VOCABULARY_WHEN_SORT=NO
igor@24 70 DONT_ADD_MARKS=NO
igor@24 71 NON_INTERACTIVE_MODE=NO
igor@25 72 PART_TO_PROCESS=''
igor@26 73 GROUP_WORDS_BY_THREE=NO
igor@26 74 GROUP_WORDS_BY_TWO=NO
igor@27 75 TAG_NAME=''
igor@27 76 MERGE_THIS_TAGS=''
igor@27 77 TAGS_LIST_ONLY=NO
igor@27 78 MERGE_TAGGED_WORDS=NO
igor@27 79 MERGE_ALL_TAGGED=NO
igor@27 80 DONT_ADD_MARKLINES=NO
igor@33 81 FILTER_WORDS=YES
igor@33 82 SHOW_VOC_STAT=NO
igor@36 83 COMPRESSED_WORDLIST=NO
igor@36 84 while getopts cl:sSkanNp:t:Tm:Mr:23 opt
igor@24 85 do
igor@24 86 case "$opt" in
igor@36 87 c) COMPRESSED_WORDLIST=YES;;
igor@24 88 s) STAT_ONLY=YES;;
igor@33 89 S) SHOW_VOC_STAT=YES;;
igor@24 90 k) NEED_TO_USE_VOCABULARY_WHEN_SORT=YES;;
igor@24 91 l) LANGUAGE="$OPTARG";;
igor@27 92 a) DONT_ADD_MARKS=YES;;
igor@24 93 n) NON_INTERACTIVE_MODE=YES;;
igor@33 94 N) FILTER_WORDS=NO;;
igor@25 95 p) PART_TO_PROCESS="$OPTARG";;
igor@27 96 t) TAG_NAME="$OPTARG";;
igor@27 97 T) TAGS_LIST_ONLY="YES";;
igor@27 98 m) DONT_ADD_MARKLINES="YES"; MERGE_TAGGED_WORDS="YES"; MERGE_THIS_TAGS="$TAG_NAME $OPTARG";;
igor@27 99 M) DONT_ADD_MARKLINES="YES"; MERGE_ALL_TAGGED="YES";;
igor@27 100 r) REMOVE_TAG="YES"; TAG_NAME="$TAG_NAME $OPTARG";;
igor@26 101 2) GROUP_WORDS_BY_TWO=YES;;
igor@26 102 3) GROUP_WORDS_BY_THREE=YES;;
igor@24 103 \?) # unknown flag
igor@24 104 show_usage
igor@24 105 exit 1;;
igor@24 106 esac
igor@24 107 done
igor@24 108 shift `expr $OPTIND - 1`
igor@24 109
igor@0 110 if [ "$1" = "-l" ]
igor@0 111 then
igor@0 112 LANGUAGE="$2"
igor@0 113 shift 2
igor@0 114 fi
igor@24 115
igor@2 116 VOCABULARY=${LANGUAGE}.txt
igor@2 117 NOTES_FILE=notes-${LANGUAGE}.txt
igor@0 118
igor@33 119 if [ "${SHOW_VOC_STAT}" = "YES" ]
igor@33 120 then
igor@33 121 $0 -l "${LANGUAGE}" -N -n ${WORK_DIR}/${VOCABULARY} | head -1 | awk '{print $5}' | tr -d "<>"
igor@33 122 exit 0
igor@33 123 fi
igor@33 124
igor@21 125 #----------------------------------------------------
igor@21 126
igor@0 127 get_words()
igor@0 128 {
igor@33 129 export FILTER_WORDS
igor@1 130 tr ' ' '\n' | sed 's/--/ /g' \
igor@3 131 | sed "s/'/__APOSTROPHE__/g" \
igor@31 132 | perl -MEncode -Mutf8 -n -e '$_ = decode( "utf8", $_);y/*\r,.:#@()+=—<>$;"?!|·[]^%&/ /; binmode STDOUT, ":utf8"; print if /^[[:alpha:] '"'"'_-]*$/'\
igor@26 133 | sed "s/__APOSTROPHE__/'/g" \
igor@3 134 | tr ' ' '\n' \
igor@3 135 | tee "$1" \
igor@3 136 | grep_v_english_perl \
igor@0 137 | sort | uniq -c | awk '{if ($2!="") print;}' | sort -rn
igor@0 138 }
igor@0 139
igor@3 140 add_stat()
igor@3 141 {
igor@27 142 if [ "$DONT_ADD_MARKLINES" = "YES" ]
igor@27 143 then
igor@27 144 cat
igor@27 145 return
igor@27 146 fi
igor@3 147 before="$1"
igor@3 148 after=${before}2
igor@3 149 cat > "$after"
igor@3 150 total="`wc -w $1 | awk '{print $1}'`"
igor@3 151 total_unknown="`cat $after|awk '{s=s+$1}END{print s}'`"
igor@3 152 total_known="`echo $total-$total_unknown|bc`"
igor@3 153 percentage="`echo '100*('$total-$total_unknown')'/$total | bc -l | sed 's/\\.\(.\).*/.\1/'`"
igor@18 154 #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@18 155 sentences="`cat $ORIGINAL_TEXT | perl -e 'local $/; $_=<>; s/[^.]//msg; print length($_);'`"
igor@18 156
igor@18 157
igor@16 158 if [ "$STAT_ONLY" = "YES" ]
igor@16 159 then
igor@18 160 echo "LANG KNOWN% UNKNOWN% KNOWN TOTAL WPS UWPS*10"
igor@18 161 echo "$LANGUAGE $percentage `echo \(100-$percentage\) | bc -l` $total_known $total `echo $total/$sentences|bc` `echo 10*$total_unknown/$sentences|bc` "
igor@17 162 rm $after
igor@16 163 return 0
igor@16 164 else
igor@34 165 groups="`echo $(grep '# groups' $after | awk '{print $3}')`"
igor@34 166 words="`echo $(grep -v '^#' $after | wc -l)`"
igor@33 167 echo "# $LANGUAGE, $percentage, <$total_known/$total>, <$groups/$words>"
igor@16 168 fi
igor@16 169
igor@14 170 PERL_SCRIPT_TEMP_NAME=`mktemp /tmp/perl-grep-v-english-XXXXXXXX`
igor@14 171 cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME
igor@14 172 my $total=shift(@ARGV);
igor@14 173 my $total_known=shift(@ARGV);
igor@14 174 my $s=0;
igor@16 175 my $mark_line=int($total_known*100/$total/5)*5;
igor@19 176 if ($mark_line>=90) {
igor@19 177 $mark_line=int($total_known*100/$total)+1;
igor@19 178 } else { $mark_line +=5; };
igor@14 179 while(<>)
igor@14 180 {
igor@33 181 next if /^#\s*groups\s*/;
igor@14 182 print;
igor@14 183 /^\s*([0-9]*)\s*/;
igor@14 184 $s+=$1;
igor@19 185 if (($total_known+$s)*100/$total>=$mark_line) {
igor@14 186 print "# $mark_line\n";
igor@14 187 if ($mark_line>=90) { $mark_line+=1; } else { $mark_line +=5; };
igor@14 188 }
igor@14 189 }
igor@14 190 PERL_SCRIPT
igor@14 191 perl $PERL_SCRIPT_TEMP_NAME "$total" "$total_known" "$after"
igor@14 192 rm $PERL_SCRIPT_TEMP_NAME
igor@3 193 rm $after
igor@3 194 }
igor@3 195
igor@26 196 two_and_three_words()
igor@26 197 {
igor@26 198 if [ $GROUP_WORDS_BY_THREE = NO -a $GROUP_WORDS_BY_TWO = NO ]
igor@26 199 then
igor@26 200 cat
igor@26 201 else
igor@26 202 cat
igor@26 203
igor@26 204 export GROUP_WORDS_BY_THREE
igor@26 205 export GROUP_WORDS_BY_TWO
igor@26 206 PERL_SCRIPT_TEMP_NAME=`mktemp /tmp/perl-two-and-three-XXXXXXXX`
igor@26 207 cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME
igor@26 208 #!/usr/bin/perl
igor@26 209 local $/;
igor@26 210 $words=<>;
igor@26 211 $words=~ s@[!?;,:#1-9".]@ @g;
igor@26 212 $words =~ s@\s+@ @g;
igor@26 213 @words = split /\s+/, $words;
igor@26 214 for ($i=0; $i<$#words-3;$i++) {
igor@26 215 my ($a,$b,$c)= ($words[$i],$words[$i+1],$words[$i+2]);
igor@26 216 if ($ENV{GROUP_WORDS_BY_THREE} eq "YES" and ($a && $b && $c)) {
igor@26 217 print "${a}_${b}_${c}\n";
igor@26 218 };
igor@26 219 if ($ENV{GROUP_WORDS_BY_TWO} eq "YES" and ($a && $b)) {
igor@26 220 print "${a}_${b}\n";
igor@26 221 };
igor@26 222 }
igor@26 223 PERL_SCRIPT
igor@26 224 perl $PERL_SCRIPT_TEMP_NAME "$ORIGINAL_TEXT"
igor@26 225 rm $PERL_SCRIPT_TEMP_NAME
igor@26 226 fi
igor@26 227 }
igor@26 228
igor@0 229 grep_v_english()
igor@0 230 {
igor@0 231 [ -e "$VOCABULARY" ] || touch "$VOCABULARY"
igor@0 232 eval $(cat $VOCABULARY | tr -d "'" | xargs -n10 echo | tr ' ' '|' | sed 's/^/egrep -xv "RRRRRRR|/' | sed 's/$/"/' | tr '\n' '|')cat
igor@0 233 }
igor@0 234
igor@0 235 grep_v_english_perl()
igor@0 236 {
igor@0 237 PERL_SCRIPT_TEMP_NAME=`mktemp /tmp/perl-grep-v-english-XXXXXXXX`
igor@0 238 cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME
igor@33 239 if ($ENV{FILTER_WORDS} eq "NO") {
igor@33 240 while(<>) { print; }
igor@33 241 exit(0);
igor@33 242 }
igor@27 243 $voc_files=$ENV{VOC_FILES};
igor@27 244 $voc_files=~s@^ @@;
igor@27 245 for $voc_file (split /\s+/,$voc_files) {
igor@27 246 if (open(VOC, $voc_file)) {
igor@27 247 while (<VOC>){
igor@27 248 chomp;
igor@27 249 #s/'//g;
igor@27 250 $voc{$_}="1";
igor@27 251 }
igor@27 252 }
igor@0 253 }
igor@0 254 while(<>) {
igor@0 255 chomp;
igor@0 256 if (not defined($voc{$_})) { print "$_\n"; }
igor@0 257 }
igor@0 258 PERL_SCRIPT
igor@0 259 [ -e "$VOCABULARY" ] || touch "$VOCABULARY"
igor@27 260 export VOCABULARY VOC_FILES
igor@27 261 VOC_FILES=$VOCABULARY
igor@27 262 for i in $TAG_NAME
igor@27 263 do
igor@27 264 VOC_FILES="${VOC_FILES} `tag_file_name $i`"
igor@27 265 done
igor@0 266 perl $PERL_SCRIPT_TEMP_NAME
igor@0 267 rm $PERL_SCRIPT_TEMP_NAME
igor@0 268 }
igor@0 269
igor@11 270 group_words()
igor@11 271 {
igor@22 272 #if [ "$LANGUAGE" != "en" ]
igor@22 273 #then
igor@22 274 # cat
igor@22 275 # return
igor@22 276 #fi
igor@11 277 PERL_SCRIPT_TEMP_NAME=`mktemp /tmp/perl-group-words-XXXXXXXX`
igor@11 278 cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME
igor@11 279 #!/usr/bin/perl
igor@11 280
igor@29 281 use Encode;
igor@29 282 use utf8;
igor@32 283 use Lingua::Stem::Snowball qw(stem);
igor@29 284
igor@12 285 eval {
igor@12 286 # http://stackoverflow.com/questions/251694/how-can-i-check-if-i-have-a-perl-module-before-using-it
igor@12 287 require String::Similarity;
igor@12 288 String::Similarity->import();
igor@12 289 };
igor@12 290 unless($@)
igor@12 291 {
igor@12 292 our $HAVE_String_Similarity=1;
igor@12 293 }
igor@12 294
igor@32 295
igor@32 296 sub load_notes_dict()
igor@32 297 {
igor@32 298 my %dict;
igor@32 299 if (open(NOTES, $ENV{NOTES_FILE})) {
igor@32 300 while(<NOTES>) {
igor@32 301 $_ = decode( "utf8", $_);
igor@32 302 chomp;
igor@32 303 s/^\s+//;
igor@32 304 my ($a,$b)=split /\s+/,$_,2;
igor@32 305 $dict{$a}=$b;
igor@32 306 }
igor@32 307 }
igor@32 308 return %dict;
igor@32 309 }
igor@32 310
igor@12 311 sub similar($$){
igor@12 312 my $a=shift;
igor@12 313 my $b=shift;
igor@12 314 if ($HAVE_String_Similarity) {
igor@12 315 return $Similarity{"$a $b"};
igor@12 316 }
igor@12 317 else {
igor@12 318 return 0;
igor@12 319 }
igor@12 320 }
igor@12 321
igor@32 322
igor@32 323 sub normalize_without_linked($)
igor@11 324 {
igor@22 325 if ( $ENV{LANGUAGE} eq "en" ) { return normalize_english(shift); }
igor@22 326 elsif ( $ENV{LANGUAGE} eq "de" ) { return normalize_german(shift); }
igor@29 327 elsif ( $ENV{LANGUAGE} eq "uk" ) { return normalize_ukrainian(shift); }
igor@31 328 elsif ( $ENV{LANGUAGE} eq "io" ) { return normalize_esperanto(shift); }
igor@22 329 else { return shift ; }
igor@22 330 }
igor@22 331
igor@32 332 sub normalize_with_linked($)
igor@32 333 {
igor@32 334 my $word = normalize_without_linked(shift);
igor@32 335 #return $word;
igor@32 336 if ($linked_words{$word}) {
igor@32 337 return $linked_words{$word};
igor@32 338 }
igor@32 339 else {
igor@32 340 return $word;
igor@32 341 }
igor@32 342 }