new-words

annotate new-words.sh @ 32:753fb84437aa

find changesets by author, revision, files, or words in the commit message
link between words support; link is created using @
author Igor Chubin <igor@chub.in>
date Fri Nov 05 20:07:46 2010 +0100 (2010-11-05)
parents 48ca8248e9cc
children 720a701b2ba9
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@27 17 -a don't add marks (and don't save marks added by user)
igor@25 18 -p pages work with specified pages only (pages = start-stop/total )
igor@20 19 -s show the text statistics (percentage of known words and so on) and exit
igor@27 20 -t tag tag known words with tag
igor@27 21 -T show list of active tags
igor@27 22 -m tag merge the words tagged with "tag" into the main vocabulary
igor@27 23 -M merge the words tagged with any tag into the main vocabulary
igor@27 24 -r tag remove subvocabulary for the "tag"
igor@26 25 -2 -3 find 2 and 3 words' sequences
igor@16 26
igor@20 27 The language of the text can be specified also
igor@20 28 by name of the program new-words (correspondent link must be created before).
igor@20 29 For example, these calls are equivalent:
igor@0 30
igor@20 31 de-words URL
igor@0 32 new-words -l de URL
igor@0 33
igor@20 34 HELP
igor@20 35 }
igor@0 36
igor@20 37 if [ "$1" = "-h" ]
igor@20 38 then
igor@20 39 show_usage
igor@20 40 exit 0
igor@20 41 fi
igor@0 42
igor@0 43 WORK_DIR=~/.new-words/
igor@28 44 TEMP1=`mktemp /tmp/new-words-temp1.XXXXXXXXXX`
igor@28 45 TEMP2=`mktemp /tmp/new-words-temp2.XXXXXXXXXX`
igor@28 46 export ORIGINAL_TEXT=`mktemp /tmp/new-words-orig.XXXXXXXXXX`
igor@0 47 editor=${EDITOR:-vim}
igor@0 48
igor@24 49 # language detection
igor@21 50
igor@2 51 LANGUAGE=en
igor@2 52 my_name="`echo $0 | sed s@.*/@@ | sed s/-.*// `"
igor@22 53 for arg
igor@22 54 do
igor@22 55 if echo "$arg" | grep -q http://...wikipedia.org/wiki/
igor@22 56 then
igor@22 57 LANGUAGE="`echo $arg | sed s@http://@@ | sed s@.wikipedia.*@@`"
igor@22 58 fi
igor@22 59 done
igor@2 60 [ "${my_name}" = "new" ] || LANGUAGE="$my_name"
igor@24 61
igor@24 62 #----------------------------------------------------
igor@24 63 # command line options processing
igor@24 64
igor@24 65 STAT_ONLY=NO
igor@24 66 NEED_TO_USE_VOCABULARY_WHEN_SORT=NO
igor@24 67 DONT_ADD_MARKS=NO
igor@24 68 NON_INTERACTIVE_MODE=NO
igor@25 69 PART_TO_PROCESS=''
igor@26 70 GROUP_WORDS_BY_THREE=NO
igor@26 71 GROUP_WORDS_BY_TWO=NO
igor@27 72 TAG_NAME=''
igor@27 73 MERGE_THIS_TAGS=''
igor@27 74 TAGS_LIST_ONLY=NO
igor@27 75 MERGE_TAGGED_WORDS=NO
igor@27 76 MERGE_ALL_TAGGED=NO
igor@27 77 DONT_ADD_MARKLINES=NO
igor@27 78 while getopts l:skanp:t:Tm:Mr:23 opt
igor@24 79 do
igor@24 80 case "$opt" in
igor@24 81 s) STAT_ONLY=YES;;
igor@24 82 k) NEED_TO_USE_VOCABULARY_WHEN_SORT=YES;;
igor@24 83 l) LANGUAGE="$OPTARG";;
igor@27 84 a) DONT_ADD_MARKS=YES;;
igor@24 85 n) NON_INTERACTIVE_MODE=YES;;
igor@25 86 p) PART_TO_PROCESS="$OPTARG";;
igor@27 87 t) TAG_NAME="$OPTARG";;
igor@27 88 T) TAGS_LIST_ONLY="YES";;
igor@27 89 m) DONT_ADD_MARKLINES="YES"; MERGE_TAGGED_WORDS="YES"; MERGE_THIS_TAGS="$TAG_NAME $OPTARG";;
igor@27 90 M) DONT_ADD_MARKLINES="YES"; MERGE_ALL_TAGGED="YES";;
igor@27 91 r) REMOVE_TAG="YES"; TAG_NAME="$TAG_NAME $OPTARG";;
igor@26 92 2) GROUP_WORDS_BY_TWO=YES;;
igor@26 93 3) GROUP_WORDS_BY_THREE=YES;;
igor@24 94 \?) # unknown flag
igor@24 95 show_usage
igor@24 96 exit 1;;
igor@24 97 esac
igor@24 98 done
igor@24 99 shift `expr $OPTIND - 1`
igor@24 100
igor@0 101 if [ "$1" = "-l" ]
igor@0 102 then
igor@0 103 LANGUAGE="$2"
igor@0 104 shift 2
igor@0 105 fi
igor@24 106
igor@2 107 VOCABULARY=${LANGUAGE}.txt
igor@2 108 NOTES_FILE=notes-${LANGUAGE}.txt
igor@0 109
igor@21 110 #----------------------------------------------------
igor@21 111
igor@0 112 get_words()
igor@0 113 {
igor@1 114 tr ' ' '\n' | sed 's/--/ /g' \
igor@3 115 | sed "s/'/__APOSTROPHE__/g" \
igor@31 116 | perl -MEncode -Mutf8 -n -e '$_ = decode( "utf8", $_);y/*\r,.:#@()+=—<>$;"?!|·[]^%&/ /; binmode STDOUT, ":utf8"; print if /^[[:alpha:] '"'"'_-]*$/'\
igor@26 117 | sed "s/__APOSTROPHE__/'/g" \
igor@3 118 | tr ' ' '\n' \
igor@3 119 | tee "$1" \
igor@3 120 | grep_v_english_perl \
igor@0 121 | sort | uniq -c | awk '{if ($2!="") print;}' | sort -rn
igor@0 122 }
igor@0 123
igor@3 124 add_stat()
igor@3 125 {
igor@27 126 if [ "$DONT_ADD_MARKLINES" = "YES" ]
igor@27 127 then
igor@27 128 cat
igor@27 129 return
igor@27 130 fi
igor@3 131 before="$1"
igor@3 132 after=${before}2
igor@3 133 cat > "$after"
igor@3 134 total="`wc -w $1 | awk '{print $1}'`"
igor@3 135 total_unknown="`cat $after|awk '{s=s+$1}END{print s}'`"
igor@3 136 total_known="`echo $total-$total_unknown|bc`"
igor@3 137 percentage="`echo '100*('$total-$total_unknown')'/$total | bc -l | sed 's/\\.\(.\).*/.\1/'`"
igor@18 138 #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@18 139 sentences="`cat $ORIGINAL_TEXT | perl -e 'local $/; $_=<>; s/[^.]//msg; print length($_);'`"
igor@18 140
igor@18 141
igor@16 142 if [ "$STAT_ONLY" = "YES" ]
igor@16 143 then
igor@18 144 echo "LANG KNOWN% UNKNOWN% KNOWN TOTAL WPS UWPS*10"
igor@18 145 echo "$LANGUAGE $percentage `echo \(100-$percentage\) | bc -l` $total_known $total `echo $total/$sentences|bc` `echo 10*$total_unknown/$sentences|bc` "
igor@17 146 rm $after
igor@16 147 return 0
igor@16 148 else
igor@16 149 echo "# $LANGUAGE, $percentage, <$total_known/$total>"
igor@16 150 fi
igor@16 151
igor@14 152 PERL_SCRIPT_TEMP_NAME=`mktemp /tmp/perl-grep-v-english-XXXXXXXX`
igor@14 153 cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME
igor@14 154 my $total=shift(@ARGV);
igor@14 155 my $total_known=shift(@ARGV);
igor@14 156 my $s=0;
igor@16 157 my $mark_line=int($total_known*100/$total/5)*5;
igor@19 158 if ($mark_line>=90) {
igor@19 159 $mark_line=int($total_known*100/$total)+1;
igor@19 160 } else { $mark_line +=5; };
igor@14 161 while(<>)
igor@14 162 {
igor@14 163 print;
igor@14 164 /^\s*([0-9]*)\s*/;
igor@14 165 $s+=$1;
igor@19 166 if (($total_known+$s)*100/$total>=$mark_line) {
igor@14 167 print "# $mark_line\n";
igor@14 168 if ($mark_line>=90) { $mark_line+=1; } else { $mark_line +=5; };
igor@14 169 }
igor@14 170 }
igor@14 171 PERL_SCRIPT
igor@14 172 perl $PERL_SCRIPT_TEMP_NAME "$total" "$total_known" "$after"
igor@14 173 rm $PERL_SCRIPT_TEMP_NAME
igor@3 174 rm $after
igor@3 175 }
igor@3 176
igor@26 177 two_and_three_words()
igor@26 178 {
igor@26 179 if [ $GROUP_WORDS_BY_THREE = NO -a $GROUP_WORDS_BY_TWO = NO ]
igor@26 180 then
igor@26 181 cat
igor@26 182 else
igor@26 183 cat
igor@26 184
igor@26 185 export GROUP_WORDS_BY_THREE
igor@26 186 export GROUP_WORDS_BY_TWO
igor@26 187 PERL_SCRIPT_TEMP_NAME=`mktemp /tmp/perl-two-and-three-XXXXXXXX`
igor@26 188 cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME
igor@26 189 #!/usr/bin/perl
igor@26 190 local $/;
igor@26 191 $words=<>;
igor@26 192 $words=~ s@[!?;,:#1-9".]@ @g;
igor@26 193 $words =~ s@\s+@ @g;
igor@26 194 @words = split /\s+/, $words;
igor@26 195 for ($i=0; $i<$#words-3;$i++) {
igor@26 196 my ($a,$b,$c)= ($words[$i],$words[$i+1],$words[$i+2]);
igor@26 197 if ($ENV{GROUP_WORDS_BY_THREE} eq "YES" and ($a && $b && $c)) {
igor@26 198 print "${a}_${b}_${c}\n";
igor@26 199 };
igor@26 200 if ($ENV{GROUP_WORDS_BY_TWO} eq "YES" and ($a && $b)) {
igor@26 201 print "${a}_${b}\n";
igor@26 202 };
igor@26 203 }
igor@26 204 PERL_SCRIPT
igor@26 205 perl $PERL_SCRIPT_TEMP_NAME "$ORIGINAL_TEXT"
igor@26 206 rm $PERL_SCRIPT_TEMP_NAME
igor@26 207 fi
igor@26 208 }
igor@26 209
igor@0 210 grep_v_english()
igor@0 211 {
igor@0 212 [ -e "$VOCABULARY" ] || touch "$VOCABULARY"
igor@0 213 eval $(cat $VOCABULARY | tr -d "'" | xargs -n10 echo | tr ' ' '|' | sed 's/^/egrep -xv "RRRRRRR|/' | sed 's/$/"/' | tr '\n' '|')cat
igor@0 214 }
igor@0 215
igor@0 216 grep_v_english_perl()
igor@0 217 {
igor@0 218 PERL_SCRIPT_TEMP_NAME=`mktemp /tmp/perl-grep-v-english-XXXXXXXX`
igor@0 219 cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME
igor@27 220 $voc_files=$ENV{VOC_FILES};
igor@27 221 $voc_files=~s@^ @@;
igor@27 222 for $voc_file (split /\s+/,$voc_files) {
igor@27 223 if (open(VOC, $voc_file)) {
igor@27 224 while (<VOC>){
igor@27 225 chomp;
igor@27 226 #s/'//g;
igor@27 227 $voc{$_}="1";
igor@27 228 }
igor@27 229 }
igor@0 230 }
igor@0 231 while(<>) {
igor@0 232 chomp;
igor@0 233 if (not defined($voc{$_})) { print "$_\n"; }
igor@0 234 }
igor@0 235 PERL_SCRIPT
igor@0 236 [ -e "$VOCABULARY" ] || touch "$VOCABULARY"
igor@27 237 export VOCABULARY VOC_FILES
igor@27 238 VOC_FILES=$VOCABULARY
igor@27 239 for i in $TAG_NAME
igor@27 240 do
igor@27 241 VOC_FILES="${VOC_FILES} `tag_file_name $i`"
igor@27 242 done
igor@0 243 perl $PERL_SCRIPT_TEMP_NAME
igor@0 244 rm $PERL_SCRIPT_TEMP_NAME
igor@0 245 }
igor@0 246
igor@11 247 group_words()
igor@11 248 {
igor@22 249 #if [ "$LANGUAGE" != "en" ]
igor@22 250 #then
igor@22 251 # cat
igor@22 252 # return
igor@22 253 #fi
igor@11 254 PERL_SCRIPT_TEMP_NAME=`mktemp /tmp/perl-group-words-XXXXXXXX`
igor@11 255 cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME
igor@11 256 #!/usr/bin/perl
igor@11 257
igor@29 258 use Encode;
igor@29 259 use utf8;
igor@32 260 use Lingua::Stem::Snowball qw(stem);
igor@29 261
igor@12 262 eval {
igor@12 263 # http://stackoverflow.com/questions/251694/how-can-i-check-if-i-have-a-perl-module-before-using-it
igor@12 264 require String::Similarity;
igor@12 265 String::Similarity->import();
igor@12 266 };
igor@12 267 unless($@)
igor@12 268 {
igor@12 269 our $HAVE_String_Similarity=1;
igor@12 270 }
igor@12 271
igor@32 272
igor@32 273 sub load_notes_dict()
igor@32 274 {
igor@32 275 my %dict;
igor@32 276 if (open(NOTES, $ENV{NOTES_FILE})) {
igor@32 277 while(<NOTES>) {
igor@32 278 $_ = decode( "utf8", $_);
igor@32 279 chomp;
igor@32 280 s/^\s+//;
igor@32 281 my ($a,$b)=split /\s+/,$_,2;
igor@32 282 $dict{$a}=$b;
igor@32 283 }
igor@32 284 }
igor@32 285 return %dict;
igor@32 286 }
igor@32 287
igor@12 288 sub similar($$){
igor@12 289 my $a=shift;
igor@12 290 my $b=shift;
igor@12 291 if ($HAVE_String_Similarity) {
igor@12 292 return $Similarity{"$a $b"};
igor@12 293 }
igor@12 294 else {
igor@12 295 return 0;
igor@12 296 }
igor@12 297 }
igor@12 298
igor@32 299
igor@32 300 sub normalize_without_linked($)
igor@11 301 {
igor@22 302 if ( $ENV{LANGUAGE} eq "en" ) { return normalize_english(shift); }
igor@22 303 elsif ( $ENV{LANGUAGE} eq "de" ) { return normalize_german(shift); }
igor@29 304 elsif ( $ENV{LANGUAGE} eq "uk" ) { return normalize_ukrainian(shift); }
igor@31 305 elsif ( $ENV{LANGUAGE} eq "io" ) { return normalize_esperanto(shift); }
igor@22 306 else { return shift ; }
igor@22 307 }
igor@22 308
igor@32 309 sub normalize_with_linked($)
igor@32 310 {
igor@32 311 my $word = normalize_without_linked(shift);
igor@32 312 #return $word;
igor@32 313 if ($linked_words{$word}) {
igor@32 314 return $linked_words{$word};
igor@32 315 }
igor@32 316 else {
igor@32 317 return $word;
igor@32 318 }
igor@32 319 }
igor@32 320
igor@32 321 sub normalize($)
igor@32 322 {
igor@32 323 return normalize_with_linked(shift);
igor@32 324 }
igor@32 325
igor@29 326 sub normalize_ukrainian($)
igor@29 327 {
igor@29 328 $_=lc(shift);
igor@29 329 s/[юіоеуаи]$//g;
igor@29 330 return $_;
igor@29 331 }
igor@29 332
igor@31 333 sub normalize_esperanto($)
igor@31 334 {
igor@31 335 $_=lc(shift);
igor@31 336 # verbs
igor@31 337 s/i$//; s/is$//; s/os$//; s/as$//; s/us$//;
igor@31 338
igor@31 339 # nouns
igor@31 340 s/j?n?$//;
igor@31 341
igor@31 342 return $_;