new-words: 4bbe553c1ee2 new-words.sh

new-words

view new-words.sh @ 12:4bbe553c1ee2

Удаление суффиксов в английских словах.

+ функция для поиска похожих слов similarity,
но она пока, хотя и присутствует в коде,
но не используется

author	Igor Chubin <igor@chub.in>
date	Sat Apr 03 00:27:00 2010 +0300 (2010-04-03)
parents	34d0332f238c
children	975b549364f2

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.*@@`"

29 fi

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

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

32 then

33 LANGUAGE="$2"

34 VOCABULARY="$LANGUAGE".txt

35 shift 2

36 fi

37 VOCABULARY=${LANGUAGE}.txt

38 NOTES_FILE=notes-${LANGUAGE}.txt

40 get_words()

41 {

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

51 }

53 add_stat()

54 {

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

65 }

67 grep_v_english()

68 {

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

71 }

73 grep_v_english_perl()

74 {

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";

83 }

84 while(<>) {

85 chomp;

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

87 }

88 PERL_SCRIPT

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

90 export VOCABULARY

91 perl $PERL_SCRIPT_TEMP_NAME

92 rm $PERL_SCRIPT_TEMP_NAME

93 }

95 group_words()

96 {

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

105

106 eval {

107 # http://stackoverflow.com/questions/251694/how-can-i-check-if-i-have-a-perl-module-before-using-it

108 require String::Similarity;

109 String::Similarity->import();

110 };

111 unless($@)

112 {

113 our $HAVE_String_Similarity=1;

114 }

115

116 sub similar($$){

117 my $a=shift;

118 my $b=shift;

119 if ($HAVE_String_Similarity) {

120 return $Similarity{"$a $b"};

121 }

122 else {

123 return 0;

124 }

125 }

126

127 sub normalize($)

128 {

129 $_=lc(shift);

130

131 s/s$//;

132 s/ed$//;

133 s/en$//;

134 s/er$//;

135 s/est$//;

136 s/ing$//;

137

138 s/ism$//; s/ist$//; s/ful$//; s/able$//; s/ably$//;

139 s/ation$//; s/ness$//; s/ally$//;

140 s/ment$//; s/ify$//; s/ity$//; s/fy$//; s/ly$//;

141 s/ise$//; s/ize$//;

142

143 s/e$//;

144 return $_;

145 }

146

147

148 sub compare($$)

149 {

150 my $a=shift;

151 my $b=shift;

152 $a =~ s/^\s*//;

153 $b =~ s/^\s*//;

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

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

156

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

158

159 if ($cmp) {

160 return $cmp;

161 }

162 else {

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

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

165 }

166 else {

167 return $a1 <=> $b1;

168 }

169 }

170 }

171

172 @lines=<>;

173 for $L (@lines) {

174 chomp($L);

175 $l=$L;

176 $l =~ s/^\s*//;

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

178 $group_weight{normalize($b)}+=$a;

179 }

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

181 for $l (@lines2) {

182 print "$l\n";

183 }

184 PERL_SCRIPT

185 perl $PERL_SCRIPT_TEMP_NAME

186 rm $PERL_SCRIPT_TEMP_NAME

187 }

188

189 text_from_url()

190 {

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

192 }

193

194 add_marks()

195 {

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

197 cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME

198 $file = $ARGV[0];

199 our $dict;

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

201 while(<NOTES>) {

202 chomp;

203 s/^\s+//;

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

205 $dict{$a}=$b;

206 }

207 }

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

209 @lines=<F>;

210 close(F);

211

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

213 for (@lines) {

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

215 $name=$1;

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

217 chomp;

218 $mark=$dict{$name};

219 $space=" "x(30-length($_));

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

221 }

222 else {

223 print F "$_";

224 }

225 }

226 close(F);

227 }

228 }

229 PERL_SCRIPT

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

231 export NOTES_FILE

232 perl $PERL_SCRIPT_TEMP_NAME "$1"

233 rm $PERL_SCRIPT_TEMP_NAME

234 }

235

236 remove_marks()

237 {

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

239 cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME

240 $file = $ARGV[0];

241 our %dict;

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

243 @lines=<F>;

244 close(F);

245

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

247 for (@lines) {

248 chomp;

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

250 my $name=$4;

251 my $comment=$6;

252 $dict{$name}=$comment;

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

254 }

255 else {

256 print F "$_\n";

257 }

258 }

259 }

260 }

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

262 @lines=<NOTES>;

263 close(NOTES);

264

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

266 for (@lines) {

267 chomp;

268 s/^\s+//;

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

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

271 print NOTES "$_\n";

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

273 }

274 }

275 for (keys %dict) {

276 $mark=$dict{$_};

277 $space=" "x(30-length($_));

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

279 }

280 }

281 }

282 PERL_SCRIPT

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

284 export NOTES_FILE

285 perl $PERL_SCRIPT_TEMP_NAME "$1"

286 rm $PERL_SCRIPT_TEMP_NAME

287 }

288

289 mkdir -p $WORK_DIR

290 oldpwd="$PWD"

291 cd $WORK_DIR

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

293 then

294 text_from_url "$1"

295 elif [ "$#" != 0 ]

296 then

297 cat "$oldpwd/$1"

298 else

299 cat

300 fi \

301 | tee $ORIGINAL_TEXT \

302 | get_words ${TEMP1}-full \

303 | group_words \

304 | add_stat ${TEMP1}-full \

305 | tee "$TEMP1" > "$TEMP2"

306

307 add_marks "$TEMP2"

308 if [ "$editor" = vim ]

309 then

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

311 else

312 echo 2

313 $editor "$TEMP2"

314 fi

315 remove_marks "$TEMP2"

316

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

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