new-words: 34d0332f238c new-words.sh

new-words

view new-words.sh @ 11:34d0332f238c

Группировка похожих слов (пока что только для английского языка).

Выполняется группировка похожих слов,
и слова сортируются не отдельно, а группами,
исходя из количества слов в группе.
В пределах группы сортировка выполняется по количеству вхождений
для каждого слова.

author	Igor Chubin <igor@chub.in>
date	Fri Apr 02 19:46:44 2010 +0300 (2010-04-02)
parents	0ff259b8b6a3
children	4bbe553c1ee2

line source

1 #!/bin/sh

2 cat <<HELP > /dev/null

4 Поддержка нескольких языков:

6 new-words -l lang URL

8 Например, для немецких текстов:

10 new-words -l de URL

12 Или, предварительно создав соответствующую ссылку:

14 de-words URL

16 HELP

18 WORK_DIR=~/.new-words/

19 TEMP1=`mktemp /tmp/news-words-XXXXXXXXXX`

20 TEMP2=`mktemp /tmp/news-words-XXXXXXXXXX`

21 export ORIGINAL_TEXT=`mktemp /tmp/news-words-XXXXXXXXXX-orig`

22 editor=${EDITOR:-vim}

24 LANGUAGE=en

25 my_name="`echo $0 | sed s@.*/@@ | sed s/-.*// `"

26 if echo "$1" | grep -q http://...wikipedia.org/wiki/

27 then

28 LANGUAGE="`echo $1 | sed s@http://@@ | sed s@.wikipedia.*@@`"

30 [ "${my_name}" = "new" ] || LANGUAGE="$my_name"

31 if [ "$1" = "-l" ]

32 then

33 LANGUAGE="$2"

34 VOCABULARY="$LANGUAGE".txt

35 shift 2

37 VOCABULARY=${LANGUAGE}.txt

38 NOTES_FILE=notes-${LANGUAGE}.txt

40 get_words()

42 tr ' ' '\n' | sed 's/--/ /g' \

43 | sed "s/'/__APOSTROPHE__/g" \

44 | tr '—·-' '-----' \

45 | tr '*\r,.:#@()+=—<>$;"?!|·[]^%&' ' ' \

46 | tr ' ' '\n' \

47 | grep -x '[[:alpha:]'"'"'-]*' \

48 | tee "$1" \

49 | grep_v_english_perl \

50 | sort | uniq -c | awk '{if ($2!="") print;}' | sort -rn

53 add_stat()

55 before="$1"

56 after=${before}2

57 cat > "$after"

58 total="`wc -w $1 | awk '{print $1}'`"

59 total_unknown="`cat $after|awk '{s=s+$1}END{print s}'`"

60 total_known="`echo $total-$total_unknown|bc`"

61 percentage="`echo '100*('$total-$total_unknown')'/$total | bc -l | sed 's/\\.\(.\).*/.\1/'`"

62 echo "# $LANGUAGE, $percentage, <$total_known/$total>"

63 cat "$after"

64 rm $after

67 grep_v_english()

69 [ -e "$VOCABULARY" ] || touch "$VOCABULARY"

70 eval $(cat $VOCABULARY | tr -d "'" | xargs -n10 echo | tr ' ' '|' | sed 's/^/egrep -xv "RRRRRRR|/' | sed 's/$/"/' | tr '\n' '|')cat

73 grep_v_english_perl()

75 PERL_SCRIPT_TEMP_NAME=`mktemp /tmp/perl-grep-v-english-XXXXXXXX`

76 cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME

77 open(VOC, $ENV{VOCABULARY})

78 or die "Can't open VOCABULARY";

79 while (<VOC>){

80 chomp;

81 #s/'//g;

82 $voc{$_}="1";

84 while(<>) {

85 chomp;

86 if (not defined($voc{$_})) { print "$_\n"; }

88 PERL_SCRIPT

89 [ -e "$VOCABULARY" ] || touch "$VOCABULARY"

90 export VOCABULARY

91 perl $PERL_SCRIPT_TEMP_NAME

92 rm $PERL_SCRIPT_TEMP_NAME

95 group_words()

97 if [ "$LANGUAGE" != "en" ]

98 then

99 cat

100 return

101 fi

102 PERL_SCRIPT_TEMP_NAME=`mktemp /tmp/perl-group-words-XXXXXXXX`

103 cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME

104 #!/usr/bin/perl

106 sub normalize($)

108 $_=lc(shift);

109 s///;

110 s/s$//;

111 s/ed$//;

112 s/ing$//;

113 return $_;

116 sub compare($$)

118 my $a=shift;

119 my $b=shift;

120 $a =~ s/^\s*//;

121 $b =~ s/^\s*//;

122 my ($a1, $a2)= split /\s+/,$a,2;

123 my ($b1, $b2)= split /\s+/,$b,2;

125 my $cmp = $group_weight{normalize($a2)} <=> $group_weight{normalize($b2)};

126 if ($cmp) {

127 return $cmp;

128 }

129 else {

130 if (normalize($a2) ne normalize($b2)) {

131 return normalize($a2) cmp normalize($b2);

132 }

133 else {

134 return $a1 <=> $b1;

135 }

136 }

139 @lines=<>;

140 for $L (@lines) {

141 chomp($L);

142 $l=$L;

143 $l =~ s/^\s*//;

144 my ($a, $b)=split(/\s+/,$l,2);

145 $group_weight{normalize($b)}+=$a;

147 @lines2 = sort { compare($b,$a) } @lines;

148 for $l (@lines2) {

149 print "$l\n";

151 PERL_SCRIPT

152 perl $PERL_SCRIPT_TEMP_NAME

153 rm $PERL_SCRIPT_TEMP_NAME

156 text_from_url()

158 lynx -dump "$1" | perl -p -e 's@http://[a-zA-Z&_.:/0-9%?=,#+()\[\]~-]*@@'

161 add_marks()

163 PERL_SCRIPT_TEMP_NAME=`mktemp /tmp/perl-grep-v-english-XXXXXXXX`

164 cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME

165 $file = $ARGV[0];

166 our $dict;

167 if (open(NOTES, $ENV{NOTES_FILE})) {

168 while(<NOTES>) {

169 chomp;

170 s/^\s+//;

171 my ($a,$b)=split /\s+/,$_,2;

172 $dict{$a}=$b;

173 }

175 if (open(F, $file)) {

176 @lines=<F>;

177 close(F);

179 if (open(F, ">$file")) {

180 for (@lines) {

181 m/\s+\S+\s+(\S+)/;

182 $name=$1;

183 if (not /^#/ and defined($dict{$name})) {

184 chomp;

185 $mark=$dict{$name};

186 $space=" "x(30-length($_));

187 print F "$_$space$mark\n";

188 }

189 else {

190 print F "$_";

191 }

192 }

193 close(F);

194 }

196 PERL_SCRIPT

197 [ -e "$NOTES_FILE" ] || touch "$NOTES_FILE"

198 export NOTES_FILE

199 perl $PERL_SCRIPT_TEMP_NAME "$1"

200 rm $PERL_SCRIPT_TEMP_NAME

203 remove_marks()

205 PERL_SCRIPT_TEMP_NAME=`mktemp /tmp/perl-grep-v-english-XXXXXXXX`

206 cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME

207 $file = $ARGV[0];

208 our %dict;

209 if (open(F, $file)) {

210 @lines=<F>;

211 close(F);

213 if (open(F, ">$file")) {

214 for (@lines) {

215 chomp;

216 if (not /^#/ and m/(\s+)(\S+)(\s+)(\S+)(\s+)(.*)/) {

217 my $name=$4;

218 my $comment=$6;

219 $dict{$name}=$comment;

220 print F "$1$2$3$4\n";

221 }

222 else {

223 print F "$_\n";

224 }

225 }

226 }

228 if (open(NOTES, $ENV{NOTES_FILE})) {

229 @lines=<NOTES>;

230 close(NOTES);

232 if (open(NOTES, ">".$ENV{NOTES_FILE})) {

233 for (@lines) {

234 chomp;

235 s/^\s+//;

236 my ($a,$b)=split /\s+/,$_,2;

237 if (not defined($dict{$a}) || ($dict{$a} eq $b)) {

238 print NOTES "$_\n";

239 if (defined($dict{$a})) { unset($dict{$a}); }

240 }

241 }

242 for (keys %dict) {

243 $mark=$dict{$_};

244 $space=" "x(30-length($_));

245 print NOTES "$_$space$mark\n";

246 }

247 }

249 PERL_SCRIPT

250 [ -e "$NOTES_FILE" ] || touch "$NOTES_FILE"

251 export NOTES_FILE

252 perl $PERL_SCRIPT_TEMP_NAME "$1"

253 rm $PERL_SCRIPT_TEMP_NAME

256 mkdir -p $WORK_DIR

257 oldpwd="$PWD"

258 cd $WORK_DIR

259 if echo "$1" | grep -q http:

260 then

261 text_from_url "$1"

262 elif [ "$#" != 0 ]

263 then

264 cat "$oldpwd/$1"

265 else

266 cat

267 fi \

268 | tee $ORIGINAL_TEXT \

269 | get_words ${TEMP1}-full \

270 | group_words \

271 | add_stat ${TEMP1}-full \

272 | tee "$TEMP1" > "$TEMP2"

274 add_marks "$TEMP2"

275 if [ "$editor" = vim ]

276 then

277 vim -c 'set keywordprg='"$LANGUAGE" -c 'set iskeyword=@,48-57,/,.,-,_,+,,,#,$,%,~,=' "$TEMP2" < /dev/tty > /dev/tty

278 else

279 echo 2

280 $editor "$TEMP2"

282 remove_marks "$TEMP2"

284 diff "$TEMP1" "$TEMP2" | awk '{print $3}' | sort -u >> "$VOCABULARY"

285 rm -f "$TEMP1" "$TEMP2" "${TEMP1}-full" "$ORIGINAL_TEXT"