igor@38: #!/bin/bash
igor@38: 
igor@38: show_usage()
igor@38: {
igor@38: cat <<HELP > /dev/stderr
igor@38: 
igor@38: USAGE: 
igor@38: 
igor@38:     new-words [ -l lang ] [ -s ] [ ARG ] 
igor@38: 
igor@38: SWITCHES: 
igor@38: 
igor@38:     -h          print this screen
igor@38:     -c          show compressed wordlist: one word per group
igor@38:     -k          put higher words that are similar to the known words (only for English)
igor@38:     -l lang     override language settings
igor@38:     -n          non-interactive mode (don't run vi)
igor@38:     -N          turn off known words filtering
igor@38:     -a          don't add marks (and don't save marks added by user)
igor@38:     -p pages    work with specified pages only (pages = start-stop/total )
igor@38:     -s          show the text statistics (percentage of known words and so on) and exit
igor@38:     -S          show your vocabulary statistics (number of words and word groups)
igor@38:     -t tag      tag known words with tag
igor@38:     -T          show list of active tags
igor@38:     -m tag      merge the words tagged with "tag" into the main vocabulary
igor@38:     -M          merge the words tagged with any tag into the main vocabulary 
igor@38:     -r tag      remove subvocabulary for the "tag"
igor@38:     -2 -3       find 2 and 3 words' sequences
igor@38: 
igor@38: The language of the text can be specified also
igor@38: by name of the program new-words (correspondent link must be created before).
igor@38: For example, these calls are equivalent:
igor@38: 
igor@38:     de-words URL
igor@38:     new-words -l de URL
igor@38: 
igor@38: HELP
igor@38: }
igor@38: 
igor@38: if [ "$1" = "-h" ]
igor@38: then
igor@38:     show_usage
igor@38:     exit 0
igor@38: fi
igor@38: 
igor@38: NEW_WORDS_PY=/home/igor/hg/new-words/new-words.py
igor@38: WORK_DIR=~/.new-words/
igor@38: TEMP1=`mktemp /tmp/new-words-temp1.XXXXXXXXXX`
igor@38: TEMP2=`mktemp /tmp/new-words-temp2.XXXXXXXXXX`
igor@38: export ORIGINAL_TEXT=`mktemp /tmp/new-words-orig.XXXXXXXXXX`
igor@38: editor=${EDITOR:-vim}
igor@38: 
igor@38: # language detection
igor@38: 
igor@38: LANGUAGE=en
igor@38: my_name="`echo $0 | sed s@.*/@@ | sed s/-.*// `"
igor@38: for arg
igor@38: do
igor@38:     if echo "$arg" | grep -q http://...wikipedia.org/wiki/
igor@38:     then
igor@38:     LANGUAGE="`echo $arg | sed s@http://@@ | sed s@.wikipedia.*@@`"
igor@38:     fi
igor@38: done
igor@38: [ "${my_name}" = "new" ] || LANGUAGE="$my_name"
igor@38: 
igor@38: #----------------------------------------------------
igor@38: # command line options processing
igor@38: 
igor@38: STAT_ONLY=NO
igor@38: NEED_TO_USE_VOCABULARY_WHEN_SORT=NO
igor@38: DONT_ADD_MARKS=NO
igor@38: NON_INTERACTIVE_MODE=NO
igor@38: PART_TO_PROCESS=''
igor@38: GROUP_WORDS_BY_THREE=NO
igor@38: GROUP_WORDS_BY_TWO=NO
igor@38: TAG_NAME=''
igor@38: MERGE_THIS_TAGS=''
igor@38: TAGS_LIST_ONLY=NO
igor@38: MERGE_TAGGED_WORDS=NO
igor@38: MERGE_ALL_TAGGED=NO
igor@38: DONT_ADD_MARKLINES=NO
igor@38: FILTER_WORDS=YES
igor@38: SHOW_VOC_STAT=NO
igor@38: COMPRESSED_WORDLIST=NO
igor@38: OLD_STYLE="NO"
igor@38: while getopts Ocl:sSkanNp:t:Tm:Mr:23 opt
igor@38: do
igor@38:     case "$opt" in
igor@38:       O)  OLD_STYLE=YES;;
igor@38:       c)  COMPRESSED_WORDLIST=YES;;
igor@38:       s)  STAT_ONLY=YES;;
igor@38:       S)  SHOW_VOC_STAT=YES;;
igor@38:       k)  NEED_TO_USE_VOCABULARY_WHEN_SORT=YES;;
igor@38:       l)  LANGUAGE="$OPTARG";;
igor@38:       a)  DONT_ADD_MARKS=YES;;
igor@38:       n)  NON_INTERACTIVE_MODE=YES;;
igor@38:       N)  FILTER_WORDS=NO;;
igor@38:       p)  PART_TO_PROCESS="$OPTARG";;
igor@38:       t)  TAG_NAME="$OPTARG";;
igor@38:       T)  TAGS_LIST_ONLY="YES";;
igor@38:       m)  DONT_ADD_MARKLINES="YES"; MERGE_TAGGED_WORDS="YES"; MERGE_THIS_TAGS="$TAG_NAME $OPTARG";;
igor@38:       M)  DONT_ADD_MARKLINES="YES"; MERGE_ALL_TAGGED="YES";;
igor@38:       r)  REMOVE_TAG="YES"; TAG_NAME="$TAG_NAME $OPTARG";;
igor@38:       2)  GROUP_WORDS_BY_TWO=YES;;
igor@38:       3)  GROUP_WORDS_BY_THREE=YES;;
igor@38:       \?)       # unknown flag
igor@38:           show_usage
igor@38:           exit 1;;
igor@38:     esac
igor@38: done
igor@38: shift `expr $OPTIND - 1`
igor@38: 
igor@38: if [ "$1" = "-l" ]
igor@38: then
igor@38:     LANGUAGE="$2"
igor@38:     shift 2
igor@38: fi
igor@38: 
igor@38: VOCABULARY=${LANGUAGE}.txt
igor@38: NOTES_FILE=notes-${LANGUAGE}.txt
igor@38: 
igor@38: if [ "${SHOW_VOC_STAT}" = "YES" ]
igor@38: then
igor@38:   $0 -l "${LANGUAGE}" -N -n ${WORK_DIR}/${VOCABULARY} | head -1 | awk '{print $5}' | tr -d "<>"
igor@38:   exit 0
igor@38: fi
igor@38: 
igor@38: #----------------------------------------------------
igor@38: 
igor@38: get_words()
igor@38: {
igor@38:    if [ "$OLD_STYLE" = NO ]
igor@38:    then 
igor@38:     $NEW_WORDS_PY -l "$LANGUAGE" -f get_words "$1"
igor@38:    else
igor@38:     get_words_OLD "$@"
igor@38:    fi
igor@38: }
igor@38: 
igor@38: get_words_OLD()
igor@38: {
igor@38:     export FILTER_WORDS
igor@38: tr ' ' '\n' | sed 's/--/ /g' \
igor@38: | sed "s/'/__APOSTROPHE__/g" \
igor@38: | perl -MEncode -Mutf8 -n -e '$_ = decode( "utf8", $_);y/*\r,.:#@()+=—<>$;"?!|·[]^%&/ /; binmode STDOUT, ":utf8"; print if /^[[:alpha:] '"'"'_-]*$/'\
igor@38: | sed "s/__APOSTROPHE__/'/g" \
igor@38: | tr ' ' '\n' \
igor@38: | tee "$1" \
igor@38: | grep_v_english_perl \
igor@38: | sort | uniq -c | awk '{if ($2!="") print;}' | sort -rn
igor@38: }
igor@38: 
igor@38: add_stat()
igor@38: {
igor@38:     if [ "$DONT_ADD_MARKLINES" = "YES" ]
igor@38:     then
igor@38:         cat
igor@38:         return
igor@38:     fi
igor@38:     before="$1"
igor@38:     after=${before}2
igor@38:     cat > "$after"
igor@38:     total="`wc -w $1 | awk '{print $1}'`"
igor@38:     total_unknown="`cat $after|awk '{s=s+$1}END{print s}'`"
igor@38:     total_known="`echo $total-$total_unknown|bc`"
igor@38:     percentage="`echo '100*('$total-$total_unknown')'/$total | bc -l | sed 's/\\.\(.\).*/.\1/'`"
igor@38:     #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@38:     sentences="`cat $ORIGINAL_TEXT | perl -e 'local $/; $_=<>; s/[^.]//msg; print length($_);'`"
igor@38: 
igor@38: 
igor@38:     if [ "$STAT_ONLY" = "YES" ]
igor@38:     then
igor@38:         echo "LANG  KNOWN%  UNKNOWN%  KNOWN     TOTAL     WPS  UWPS*10"
igor@38:         echo "$LANGUAGE    $percentage    `echo \(100-$percentage\) | bc -l`      $total_known    $total    `echo $total/$sentences|bc`   `echo 10*$total_unknown/$sentences|bc` "
igor@38:         rm $after
igor@38:         return 0
igor@38:     else 
igor@38:         groups="`echo $(grep '# groups' $after | awk '{print $3}')`"
igor@38:         words="`echo $(grep -v '^#' $after | wc -l)`"
igor@38:         echo "# $LANGUAGE, $percentage, <$total_known/$total>, <$groups/$words>"
igor@38:     fi
igor@38: 
igor@38:     PERL_SCRIPT_TEMP_NAME=`mktemp /tmp/perl-grep-v-english-XXXXXXXX`
igor@38:     cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME
igor@38: my $total=shift(@ARGV);
igor@38: my $total_known=shift(@ARGV);
igor@38: my $s=0;
igor@38: my $mark_line=int($total_known*100/$total/5)*5;
igor@38: if ($mark_line>=90) { 
igor@38:     $mark_line=int($total_known*100/$total)+1;
igor@38: } else { $mark_line +=5; };
igor@38: while(<>)
igor@38: {
igor@38:     next if /^#\s*groups\s*/;
igor@38:     print;
igor@38:     /^\s*([0-9]*)\s*/;
igor@38:     $s+=$1;
igor@38:     if (($total_known+$s)*100/$total>=$mark_line) {
igor@38:         print "# $mark_line\n";
igor@38:         if ($mark_line>=90) { $mark_line+=1; } else { $mark_line +=5; };
igor@38:     }
igor@38: }
igor@38: PERL_SCRIPT
igor@38:     perl $PERL_SCRIPT_TEMP_NAME "$total" "$total_known" "$after"
igor@38:     rm $PERL_SCRIPT_TEMP_NAME
igor@38:     rm $after
igor@38: }
igor@38: 
igor@38: two_and_three_words()
igor@38: {
igor@38:     if [ $GROUP_WORDS_BY_THREE = NO -a $GROUP_WORDS_BY_TWO = NO ]
igor@38:     then 
igor@38:         cat
igor@38:     else
igor@38:         cat 
igor@38: 
igor@38:     export GROUP_WORDS_BY_THREE
igor@38:     export GROUP_WORDS_BY_TWO
igor@38:     PERL_SCRIPT_TEMP_NAME=`mktemp /tmp/perl-two-and-three-XXXXXXXX`
igor@38:     cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME
igor@38: #!/usr/bin/perl
igor@38: local $/;
igor@38: $words=<>;
igor@38: $words=~ s@[!?;,:#1-9".]@ @g;
igor@38: $words =~ s@\s+@ @g;
igor@38: @words = split /\s+/, $words;
igor@38: for ($i=0; $i<$#words-3;$i++) {
igor@38:     my ($a,$b,$c)= ($words[$i],$words[$i+1],$words[$i+2]);
igor@38:     if ($ENV{GROUP_WORDS_BY_THREE} eq "YES" and ($a && $b && $c)) {
igor@38:         print "${a}_${b}_${c}\n";
igor@38:     };  
igor@38:     if ($ENV{GROUP_WORDS_BY_TWO} eq "YES" and ($a && $b)) {
igor@38:         print "${a}_${b}\n";
igor@38:     };
igor@38: }
igor@38: PERL_SCRIPT
igor@38:     perl $PERL_SCRIPT_TEMP_NAME "$ORIGINAL_TEXT"
igor@38:     rm $PERL_SCRIPT_TEMP_NAME
igor@38:     fi
igor@38: }
igor@38: 
igor@38: grep_v_english()
igor@38: {
igor@38: [ -e "$VOCABULARY" ] || touch "$VOCABULARY"
igor@38: eval $(cat $VOCABULARY | tr -d "'" | xargs -n10 echo | tr ' ' '|' | sed 's/^/egrep -xv "RRRRRRR|/' | sed 's/$/"/' | tr '\n' '|')cat
igor@38: }
igor@38: 
igor@38: grep_v_english_perl()
igor@38: {
igor@38:     PERL_SCRIPT_TEMP_NAME=`mktemp /tmp/perl-grep-v-english-XXXXXXXX`
igor@38:     cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME
igor@38:     if ($ENV{FILTER_WORDS} eq "NO") {
igor@38:         while(<>) { print; }
igor@38:         exit(0);
igor@38:     }
igor@38: $voc_files=$ENV{VOC_FILES};
igor@38: $voc_files=~s@^ @@;
igor@38: for $voc_file (split /\s+/,$voc_files) {
igor@38:     if (open(VOC, $voc_file)) {
igor@38:         while (<VOC>){
igor@38:             chomp;
igor@38:             #s/'//g;
igor@38:             $voc{$_}="1";
igor@38:         }
igor@38:     }
igor@38: }
igor@38: while(<>) {
igor@38:     chomp;
igor@38:     if (not defined($voc{$_})) { print "$_\n"; }
igor@38: }
igor@38: PERL_SCRIPT
igor@38:     [ -e "$VOCABULARY" ] || touch "$VOCABULARY"
igor@38:     export VOCABULARY VOC_FILES
igor@38:     VOC_FILES=$VOCABULARY
igor@38:     for i in $TAG_NAME
igor@38:     do
igor@38:         VOC_FILES="${VOC_FILES} `tag_file_name $i`"
igor@38:     done
igor@38:     perl $PERL_SCRIPT_TEMP_NAME
igor@38:     rm $PERL_SCRIPT_TEMP_NAME
igor@38: }
igor@38: 
igor@38: group_words()
igor@38: {
igor@38:    if [ "$OLD_STYLE" = NO ]
igor@38:    then 
igor@38:     $NEW_WORDS_PY -l "$LANGUAGE" -f group_words "$1"
igor@38:    else
igor@38:     group_words_OLD "$@"
igor@38:    fi
igor@38: }
igor@38: 
igor@38: group_words_OLD()
igor@38: {
igor@38:     #if [ "$LANGUAGE" != "en" ]
igor@38:     #then
igor@38:     #    cat 
igor@38:     #    return
igor@38:     #fi
igor@38:     PERL_SCRIPT_TEMP_NAME=`mktemp /tmp/perl-group-words-XXXXXXXX`
igor@38:     cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME
igor@38: #!/usr/bin/perl
igor@38: 
igor@38: use Encode;
igor@38: use utf8;
igor@38: use Lingua::Stem::Snowball qw(stem);
igor@38: 
igor@38: eval {
igor@38: # http://stackoverflow.com/questions/251694/how-can-i-check-if-i-have-a-perl-module-before-using-it
igor@38:     require String::Similarity;
igor@38:     String::Similarity->import();
igor@38: };
igor@38: unless($@)
igor@38: {
igor@38:     our $HAVE_String_Similarity=1;
igor@38: }
igor@38: 
igor@38: 
igor@38: sub load_notes_dict()
igor@38: {
igor@38:     my %dict;
igor@38:     if (open(NOTES, $ENV{NOTES_FILE})) {
igor@38:         while(<NOTES>) {
igor@38:             $_ = decode( "utf8", $_);
igor@38:             chomp;
igor@38:             s/^\s+//;
igor@38:             my ($a,$b)=split /\s+/,$_,2;
igor@38:             $dict{$a}=$b;
igor@38:         }
igor@38:     }
igor@38:     return %dict;
igor@38: }
igor@38: 
igor@38: sub similar($$){
igor@38:     my $a=shift;
igor@38:     my $b=shift;
igor@38:     if ($HAVE_String_Similarity) {
igor@38:         return $Similarity{"$a $b"};
igor@38:     } 
igor@38:     else {
igor@38:         return 0;
igor@38:     }
igor@38: }
igor@38: 
igor@38: 
igor@38: sub normalize_without_linked($)
igor@38: {
igor@38:     if   ( $ENV{LANGUAGE} eq "en" ) { return normalize_english(shift); }
igor@38:     elsif ( $ENV{LANGUAGE} eq "de" ) { return normalize_german(shift); }
igor@38:     elsif ( $ENV{LANGUAGE} eq "uk" ) { return normalize_ukrainian(shift); }
igor@38:     elsif ( $ENV{LANGUAGE} eq "io" ) { return normalize_esperanto(shift); }
igor@38:     else { return shift ; }
igor@38: }
igor@38: 
igor@38: sub normalize_with_linked($)
igor@38: {
igor@38:     my $word = normalize_without_linked(shift);
igor@38:         #return $word;
igor@38:     if ($linked_words{$word}) {
igor@38:         return $linked_words{$word};
igor@38:     }
igor@38:     else {
igor@38:         return $word;
igor@38:     }
igor@38: }
igor@38: 
igor@38: sub normalize($)
igor@38: {
igor@38:     return normalize_with_linked(shift);
igor@38: }
igor@38: 
igor@38: sub normalize_ukrainian($)
igor@38: {
igor@38:     $_=lc(shift);
igor@38:     s/[юіоеуаи]$//g;
igor@38:     return $_;
igor@38: }
igor@38: 
igor@38: sub normalize_esperanto($)
igor@38: {
igor@38:     $_=lc(shift);
igor@38: # verbs
igor@38:     s/i$//; s/is$//; s/os$//; s/as$//; s/us$//;
igor@38: 
igor@38: # nouns
igor@38:     s/j?n?$//;
igor@38: 
igor@38:     return $_;
igor@38: }
igor@38: 
igor@38: sub normalize_german($)
igor@38: {
igor@38:     @stems = stem('de', \@_);
igor@38:     return $stems[0];
igor@38: }
igor@38: 
igor@38: sub normalize_german_($)
igor@38: {
igor@38:     $_=lc(shift);
igor@38: 
igor@38:     s/heit$//;  s/keit$//; s/tum$//; s/ung$//; s/nis$//;s/schaft$//; s/ist$//; 
igor@38:     s/en$//; s/er$//;
igor@38: 
igor@38:     s/lich$//; s/ig$//;
igor@38:     s/al$//; s/isch$//;
igor@38:     s/ell$//; s/haft$//;
igor@38: 
igor@38:     s/bar$//; s/sam$//; s/lich$//;
igor@38: 
igor@38:     @prefixes=qw(
igor@38:         ab an auf aus bei dazwischen ein fest heraus her hinaus hin los mit nach voraus vorbei vor weg weiter zurück zusammen zu
igor@38:         be emp ent er ge miss ver zer durch über um unter wieder);
igor@38:     @prefixes=();
igor@38:     for $pref (@prefixes) {
igor@38:         s/^$pref//;
igor@38:     }
igor@38: 
igor@38: 
igor@38:     return $_;
igor@38: }
igor@38: 
igor@38: sub normalize_english($)
igor@38: {
igor@38:     $_=lc(shift);
igor@38: 
igor@38:     s/s$//;
igor@38: 
igor@38:     s/ation$//;  s/ness$//; s/ship$//; s/ally$//; s/ance$//;s/ity$//; s/ment$//; 
igor@38: 
igor@38:     s/ed$//;
igor@38:     s/en$//;
igor@38:     s/er$//;
igor@38:     s/est$//;
igor@38:     s/ing$//;
igor@38: 
igor@38:     s/ism$//; s/ist$//; s/ful$//; s/able$//; s/ably$//;
igor@38:     s/ify$//; s/fy$//; s/ly$//;
igor@38:     s/ise$//; s/ize$//;
igor@38: 
igor@38:     s/e$//;
igor@38:     return $_;
igor@38: }
igor@38: 
igor@38: 
igor@38: sub compare($$)
igor@38: {
igor@38:     my $a=shift;
igor@38:     my $b=shift;
igor@38:     $a =~ s/^\s*//;
igor@38:     $b =~ s/^\s*//;
igor@38:     my ($a1, $a2)= split /\s+/,$a,2;
igor@38:     my ($b1, $b2)= split /\s+/,$b,2;
igor@38: 
igor@38:     my $cmp = $group_weight{normalize($a2)} <=> $group_weight{normalize($b2)};
igor@38: 
igor@38:     if ($cmp) {
igor@38:         return $cmp;
igor@38:     }
igor@38:     else {
igor@38:         if (normalize($a2) ne normalize($b2)) {
igor@38:             return normalize($a2) cmp normalize($b2);
igor@38:         }
igor@38:         else {
igor@38:             return $a1 <=> $b1;
igor@38:         }
igor@38:     }
igor@38: }
igor@38: 
igor@38: sub log_($)
igor@38: {
igor@38:     return;
igor@38:     open(LOG, ">>", "/tmp/log1");
igor@38:     print LOG $_[0];
igor@38:     close(LOG);
igor@38: }
igor@38: 
igor@38: sub find_linked_words($)
igor@38: {
igor@38:     my %linked_words;
igor@38:     my $dict = shift;
igor@38:     log_("1");
igor@38:     log_(join(" ", keys(%$dict)));
igor@38: 
igor@38:     for $key (keys(%$dict)) {
igor@38:         $val = $dict->{$key};
igor@38:         log_($key."\n");
igor@38:         if ($val =~ /\@([a-z]*)/) {
igor@38:             $linked_words{normalize($key)} = normalize($1);
igor@38:             log_(normalize($key)." = ".normalize($1)."\n");
igor@38:         }
igor@38:     }
igor@38:     return %linked_words;
igor@38: }
igor@38: 
igor@38: sub lc_length($)
igor@38: {
igor@38:     my $a= shift;
igor@38:     $a =~ s/[a-z]//g;
igor@38:     return length($a);
igor@38: }
igor@38: 
igor@38: our %dict = load_notes_dict();
igor@38: our %linked_words = find_linked_words(\%dict);
igor@38: 
igor@38: our %Vocabulary;
igor@38: open(VOC, $ENV{VOCABULARY})
igor@38:  or die "Can't open VOCABULARY";
igor@38: while (<VOC>){
igor@38:     chomp;
igor@38:     #s/'//g;
igor@38:     $Vocabulary{normalize($_)}="1";
igor@38: }
igor@38: close(VOC);
igor@38: 
igor@38: binmode STDIN,":utf8";
igor@38: @lines=<STDIN>;
igor@38: for $L (@lines) {
igor@38:     chomp($L);
igor@38:     #$L = decode( "utf8", $L);
igor@38:     $l=$L;
igor@38:     $l =~ s/^\s*//;
igor@38:     my ($a, $b)=split(/\s+/,$l,2);
igor@38:     $group_weight{normalize($b)}+=$a;
igor@38: }
igor@38: if ($ENV{NEED_TO_USE_VOCABULARY_WHEN_SORT} eq "YES") {
igor@38:     for $k (keys %group_weight) {
igor@38:         if (defined($Vocabulary{$k})) {
igor@38:             $group_weight{$k} *= 2;
igor@38:         }
igor@38:     }
igor@38: }
igor@38: @lines2 = sort { compare($b,$a) } @lines;
igor@38: binmode STDOUT, ":utf8";
igor@38: print "# groups ".scalar(keys(%group_weight))."\n";
igor@38: if ($ENV{COMPRESSED_WORDLIST} eq "YES") {
igor@38:     my $sum = 0;
igor@38:     my $min = 9999;
igor@38:     for $L (@lines2) {
igor@38:         chomp($L); $l=$L; $l =~ s/^(\s*)//; my $spaces = $1; my ($a, $b)=split(/\s+/,$l,2);
igor@38:         $group_name = normalize($b);
igor@38:         if ($group_name ne $prev_group_name and $prev_group_name ne '' ) {
igor@38:             #print (" "x(7-length($sum))),"$sum $main_word\n";
igor@38:             print +(" "x(7-length($sum))),"$sum $main_word\n";
igor@38:             $sum = $a;
igor@38:             $min = length($b) + 2*lc_length($b);
igor@38:             $main_word = $b;
igor@38:         }
igor@38:         else {
igor@38:             $sum += $a;
igor@38:             if ($min > length($b) + 2*lc_length($b)) {
igor@38:                 $min = length($b) + 2*lc_length($b);
igor@38:                 $main_word = $b;
igor@38:             }
igor@38:         }
igor@38:         $prev_group_name = $group_name;
igor@38:     }
igor@38: }
igor@38: else {
igor@38:     for $l (@lines2) {
igor@38:         print "$l\n";
igor@38:     }
igor@38: }
igor@38: PERL_SCRIPT
igor@38:     export VOCABULARY
igor@38:     export NEED_TO_USE_VOCABULARY_WHEN_SORT
igor@38:     export LANGUAGE
igor@38:     export COMPRESSED_WORDLIST
igor@38:     [ -e "$NOTES_FILE" ] || touch "$NOTES_FILE"
igor@38:     export NOTES_FILE
igor@38:     perl $PERL_SCRIPT_TEMP_NAME
igor@38:     rm $PERL_SCRIPT_TEMP_NAME
igor@38: }
igor@38: 
igor@38: text_from_url()
igor@38: {
igor@38: lynx -dump "$1" | perl -p -e 's@http://[a-zA-Z&_.:/0-9%?=,#+()\[\]~-]*@@'
igor@38: }
igor@38: 
igor@38: add_marks()
igor@38: {
igor@38:     PERL_SCRIPT_TEMP_NAME=`mktemp /tmp/perl-grep-v-english-XXXXXXXX`
igor@38:     cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME
igor@38: use Encode;
igor@38: 
igor@38: sub load_notes_dict()
igor@38: {
igor@38:     my %dict;
igor@38:     if (open(NOTES, $ENV{NOTES_FILE})) {
igor@38:         while(<NOTES>) {
igor@38:             $_ = decode( "utf8", $_);
igor@38:             chomp;
igor@38:             s/^\s+//;
igor@38:             my ($a,$b)=split /\s+/,$_,2;
igor@38:             $dict{$a}=$b;
igor@38:         }
igor@38:     }
igor@38:     return %dict;
igor@38: }
igor@38: 
igor@38: %dict = load_notes_dict();
igor@38: 
igor@38: $file = $ARGV[0];
igor@38: if (open(F, $file)) {
igor@38:     @lines=<F>;
igor@38:     close(F);
igor@38:     for (@lines) {$_ = decode( "utf8", $_);};
igor@38: 
igor@38:     if (open(F, ">$file")) {
igor@38:         binmode F, ":utf8";
igor@38:         for (@lines) {
igor@38:             m/\s+\S+\s+(\S+)/;
igor@38:             $name=$1;
igor@38:             if (not /^#/ and defined($dict{$name})) {
igor@38:                 chomp;
igor@38:                 $mark=$dict{$name};
igor@38:                 $space=" "x(30-length($_));
igor@38:                 print F "$_$space$mark\n";
igor@38:             }
igor@38:             else {
igor@38:                 print F "$_";
igor@38:             }
igor@38:         }
igor@38:         close(F);
igor@38:     }
igor@38: }
igor@38: PERL_SCRIPT
igor@38:     [ -e "$NOTES_FILE" ] || touch "$NOTES_FILE"
igor@38:     export NOTES_FILE
igor@38:     perl $PERL_SCRIPT_TEMP_NAME "$1"
igor@38:     rm $PERL_SCRIPT_TEMP_NAME
igor@38: }
igor@38: 
igor@38: remove_marks()
igor@38: {
igor@38:     PERL_SCRIPT_TEMP_NAME=`mktemp /tmp/perl-grep-v-english-XXXXXXXX`
igor@38:     cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME
igor@38: $file = $ARGV[0];
igor@38: our %dict;
igor@38: if (open(F, $file)) {
igor@38:     @lines=<F>;
igor@38:     close(F);
igor@38: 
igor@38:     if (open(F, ">$file")) {
igor@38:         for (@lines) {
igor@38:             chomp;
igor@38:             if (not /^#/ and m/(\s+)(\S+)(\s+)(\S+)(\s+)(.*)/) {
igor@38:                 my $name=$4;
igor@38:                 my $comment=$6;
igor@38:                 $dict{$name}=$comment;
igor@38:                 print F "$1$2$3$4\n";
igor@38:             }
igor@38:             else {
igor@38:                 print F "$_\n";
igor@38:             }
igor@38:         }
igor@38:     }
igor@38: }
igor@38: if (($ENV{DONT_ADD_MARKS} ne "YES") and open(NOTES, $ENV{NOTES_FILE})) {
igor@38:     @lines=<NOTES>;
igor@38:     close(NOTES);
igor@38: 
igor@38:     if (open(NOTES, ">".$ENV{NOTES_FILE})) {
igor@38:         for (@lines) {
igor@38:             chomp;
igor@38:             s/^\s+//;
igor@38:             my ($a,$b)=split /\s+/,$_,2;
igor@38:             if (not defined($dict{$a}) || ($dict{$a} eq $b)) {
igor@38:                 print NOTES "$_\n";
igor@38:                 if (defined($dict{$a})) { unset($dict{$a}); }
igor@38:             }
igor@38:         }
igor@38:         for (keys %dict) {
igor@38:             $mark=$dict{$_};
igor@38:             $space=" "x(30-length($_));
igor@38:             print NOTES "$_$space$mark\n";
igor@38:         }
igor@38:     }
igor@38: }
igor@38: PERL_SCRIPT
igor@38:     [ -e "$NOTES_FILE" ] || touch "$NOTES_FILE"
igor@38:     export NOTES_FILE
igor@38:     export DONT_ADD_MARKS
igor@38:     perl $PERL_SCRIPT_TEMP_NAME "$1"
igor@38:     rm $PERL_SCRIPT_TEMP_NAME
igor@38: }
igor@38: 
igor@38: part()
igor@38: {
igor@38:     PERL_SCRIPT_TEMP_NAME=`mktemp /tmp/perl-part-XXXXXXXX`
igor@38:     cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME
igor@38: #!/usr/bin/perl
igor@38: 
igor@38: my @lines=<STDIN>;
igor@38: my $lines=$#lines;
igor@38: my $interval=$ARGV[0];
igor@38: if (not $interval) {
igor@38:     print @lines;
igor@38: }
igor@38: else {
igor@38:     my ($start,$stop,$total);
igor@38:     if ($interval =~ m@(.*)/(.*)@) {
igor@38:         $start = $1;
igor@38:         $total = $2;
igor@38:     }
igor@38:     else {
igor@38:         $start=$interval;
igor@38:         $total=0;
igor@38:     }
igor@38:     if ($start =~ m@(.*)-(.*)@) {
igor@38:         $start = $1;
igor@38:         $stop = $2;
igor@38:     }
igor@38:     if ($start =~ m@(.*)\+(.*)@) {
igor@38:         $start = $1;
igor@38:         $stop = $start+$2;
igor@38:     }
igor@38: 
igor@38:     $start=int($lines/$total*$start);
igor@38:     $stop=int($lines/$total*$stop);
igor@38: 
igor@38:     for($i=$start;$i<$stop;$i++){
igor@38:         print $lines[$i];
igor@38:     }
igor@38: }
igor@38: PERL_SCRIPT
igor@38:     perl $PERL_SCRIPT_TEMP_NAME "$1"
igor@38:     rm $PERL_SCRIPT_TEMP_NAME
igor@38: }
igor@38: 
igor@38: if [ "$TAGS_LIST_ONLY" = "YES" ] 
igor@38: then
igor@38:     cd "${WORK_DIR}"
igor@38:     echo ${LANGUAGE}_*.txt | tr ' ' '\n'  | grep -v '*' | sed 's/[^_]*_//;s/.txt$//'
igor@38:     exit 0
igor@38: fi
igor@38: 
igor@38: tag_file_name()
igor@38: {
igor@38:     echo "${LANGUAGE}_${1}.txt"
igor@38: }
igor@38: 
igor@38: if [ "$REMOVE_TAG" = "YES" ]
igor@38: then
igor@38:     cd "${WORK_DIR}"
igor@38:     for i in $TAG_NAME 
igor@38:     do
igor@38:         echo "$TAGNAME" | grep -q '[/*?]' && continue
igor@38:         f="`tag_file_name $i`"
igor@38:         if [ -e "$f" ] 
igor@38:         then
igor@38:             rm -f "$f" && echo Tag "'$i'" removed
igor@38:         else
igor@38:             echo Unknown tag "'$i'"
igor@38:         fi
igor@38:     done
igor@38:     exit 0
igor@38: fi
igor@38: 
igor@38: mkdir -p $WORK_DIR
igor@38: oldpwd="$PWD"
igor@38: cd $WORK_DIR
igor@38: if [ "$MERGE_TAGGED_WORDS" = "YES" ]
igor@38: then
igor@38:     VOC_FILES=''
igor@38:     for i in $MERGE_THIS_TAGS
igor@38:     do
igor@38:         f=`tag_file_name $i`
igor@38:         [ -e "$f" ] && VOC_FILES="${VOC_FILES} $f"
igor@38:     done
igor@38:     if [ -z "$VOC_FILES" ]
igor@38:     then 
igor@38:         echo Unknown tags: $MERGE_THIS_TAGS > /dev/stderr
igor@38:     else
igor@38:         cat $VOC_FILES
igor@38:     fi
igor@38: elif [ "$MERGE_ALL_TAGGED" = "YES" ]
igor@38: then
igor@38:     cat ${LANGUAGE}_*.txt
igor@38: elif echo "$1" | grep -q http: 
igor@38: then 
igor@38:     text_from_url "$1"
igor@38: elif [ "$#" != 0 ]
igor@38: then
igor@38:     if echo $1 | grep -q ^/
igor@38:     then
igor@38:         cat "$1"
igor@38:     else
igor@38:         cat "$oldpwd/$1"
igor@38:     fi
igor@38: else 
igor@38:     cat
igor@38: fi \
igor@38:    | part $PART_TO_PROCESS \
igor@38:    | tee $ORIGINAL_TEXT \
igor@38:    | two_and_three_words \
igor@38:    | get_words ${TEMP1}-full \
igor@38:    | group_words \
igor@38:    | add_stat ${TEMP1}-full \
igor@38:    | tee "$TEMP1" > "$TEMP2"
igor@38: 
igor@38: if [ "$STAT_ONLY" = "YES" ]
igor@38: then
igor@38:     cat "$TEMP1"
igor@38: elif [ "$NON_INTERACTIVE_MODE" = "YES" ]
igor@38: then
igor@38:     cat "$TEMP1"
igor@38: else
igor@38:     if [ `wc -l "$TEMP2" | awk '{print $1}'` != 0 ] 
igor@38:     then
igor@38:         [ "$DONT_ADD_MARKS" = "YES" ] || add_marks "$TEMP2"
igor@38:         if [ "$editor" = vim ]
igor@38:         then
igor@38:             vim -c 'set keywordprg='"$LANGUAGE" -c 'set iskeyword=@,48-57,/,.,-,_,+,,,#,$,%,~,=,48-255' "$TEMP2" < /dev/tty > /dev/tty
igor@38:         else
igor@38:             $editor "$TEMP2"
igor@38:         fi
igor@38:         remove_marks "$TEMP2"
igor@38: 
igor@38:         vocabulary="$VOCABULARY"
igor@38:         [ -n "$TAG_NAME" ] && vocabulary="`tag_file_name $TAG_NAME`"
igor@38:         diff "$TEMP1" "$TEMP2" | awk '{print $3}' | sort -u >> "$vocabulary"
igor@38:     fi
igor@38: fi
igor@38: 
igor@38: rm -f "$TEMP1" "$TEMP2" "${TEMP1}-full" "$ORIGINAL_TEXT"