new-words: c65ffd60cc18 new-words.sh

new-words

view new-words.sh @ 16:c65ffd60cc18

ключ -s: Информация об известных словах в тексте

Может использоваться в пакетном режиме,
например, для того чтобы из множества файлов
выбрать тот, в котором процент известных слов наибольший

author	Igor Chubin <igor@chub.in>
date	Sun Apr 04 19:03:30 2010 +0300 (2010-04-04)
parents	9b18c7efe31c
children	35eeaf2620ce

line source

1 #!/bin/sh

2 cat <<HELP > /dev/null

4 USAGE:

6 new-words [ -l lang ] [ -s ] [ ARG ]

8 SWITCHES:

10 -s show text statistics and exit

11 -l lang override language settings

13 Поддержка нескольких языков:

15 new-words -l lang URL

17 Например, для немецких текстов:

19 new-words -l de URL

21 Или, предварительно создав соответствующую ссылку:

23 de-words URL

25 HELP

27 WORK_DIR=~/.new-words/

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

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

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

31 editor=${EDITOR:-vim}

33 LANGUAGE=en

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

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

36 then

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

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

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

41 then

42 LANGUAGE="$2"

43 VOCABULARY="$LANGUAGE".txt

44 shift 2

46 VOCABULARY=${LANGUAGE}.txt

47 NOTES_FILE=notes-${LANGUAGE}.txt

49 STAT_ONLY=NO

50 if [ "$1" = "-s" ]

51 then

52 STAT_ONLY=YES

53 shift

56 get_words()

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

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

60 | tr '—·-' '-----' \

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

62 | tr ' ' '\n' \

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

64 | tee "$1" \

65 | grep_v_english_perl \

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

69 add_stat()

71 before="$1"

72 after=${before}2

73 cat > "$after"

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

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

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

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

78 if [ "$STAT_ONLY" = "YES" ]

79 then

80 echo "LANG KNOWN% UNKNOWN% KNOWN TOTAL"

81 echo "$LANGUAGE $percentage `echo \(100-$percentage\) | bc -l` $total_known $total"

82 return 0

83 else

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

85 fi

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

88 cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME

89 my $total=shift(@ARGV);

90 my $total_known=shift(@ARGV);

91 my $s=0;

92 my $mark_line=int($total_known*100/$total/5)*5;

93 if ($mark_line>=90) { $mark_line+=1; } else { $mark_line +=5; };

94 while(<>)

96 print;

97 /^\s*([0-9]*)\s*/;

98 $s+=$1;

99 if (int(($total_known+$s)*100/$total)>=$mark_line) {

100 print "# $mark_line\n";

101 if ($mark_line>=90) { $mark_line+=1; } else { $mark_line +=5; };

102 }

104 PERL_SCRIPT

105 perl $PERL_SCRIPT_TEMP_NAME "$total" "$total_known" "$after"

106 rm $PERL_SCRIPT_TEMP_NAME

107 rm $after

110 grep_v_english()

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

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

116 grep_v_english_perl()

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

119 cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME

120 open(VOC, $ENV{VOCABULARY})

121 or die "Can't open VOCABULARY";

122 while (<VOC>){

123 chomp;

124 #s/'//g;

125 $voc{$_}="1";

127 while(<>) {

128 chomp;

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

131 PERL_SCRIPT

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

133 export VOCABULARY

134 perl $PERL_SCRIPT_TEMP_NAME

135 rm $PERL_SCRIPT_TEMP_NAME

138 group_words()

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

141 then

142 cat

143 return

144 fi

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

146 cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME

147 #!/usr/bin/perl

149 eval {

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

151 require String::Similarity;

152 String::Similarity->import();

154 unless($@)

156 our $HAVE_String_Similarity=1;

159 sub similar($$){

160 my $a=shift;

161 my $b=shift;

162 if ($HAVE_String_Similarity) {

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

164 }

165 else {

166 return 0;

167 }

170 sub normalize($)

172 $_=lc(shift);

174 s/s$//;

175 s/ed$//;

176 s/en$//;

177 s/er$//;

178 s/est$//;

179 s/ing$//;

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

182 s/ation$//; s/ness$//; s/ship$//; s/ally$//;

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

184 s/ise$//; s/ize$//;

186 s/e$//;

187 return $_;

191 sub compare($$)

193 my $a=shift;

194 my $b=shift;

195 $a =~ s/^\s*//;

196 $b =~ s/^\s*//;

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

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

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

202 if ($cmp) {

203 return $cmp;

204 }

205 else {

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

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

208 }

209 else {

210 return $a1 <=> $b1;

211 }

212 }

215 @lines=<>;

216 for $L (@lines) {

217 chomp($L);

218 $l=$L;

219 $l =~ s/^\s*//;

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

221 $group_weight{normalize($b)}+=$a;

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

224 for $l (@lines2) {

225 print "$l\n";

227 PERL_SCRIPT

228 perl $PERL_SCRIPT_TEMP_NAME

229 rm $PERL_SCRIPT_TEMP_NAME

232 text_from_url()

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

237 add_marks()

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

240 cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME

241 $file = $ARGV[0];

242 our $dict;

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

244 while(<NOTES>) {

245 chomp;

246 s/^\s+//;

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

248 $dict{$a}=$b;

249 }

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

252 @lines=<F>;

253 close(F);

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

256 for (@lines) {

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

258 $name=$1;

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

260 chomp;

261 $mark=$dict{$name};

262 $space=" "x(30-length($_));

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

264 }

265 else {

266 print F "$_";

267 }

268 }

269 close(F);

270 }

272 PERL_SCRIPT

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

274 export NOTES_FILE

275 perl $PERL_SCRIPT_TEMP_NAME "$1"

276 rm $PERL_SCRIPT_TEMP_NAME

279 remove_marks()

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

282 cat <<'PERL_SCRIPT' > $PERL_SCRIPT_TEMP_NAME

283 $file = $ARGV[0];

284 our %dict;

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

286 @lines=<F>;

287 close(F);

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

290 for (@lines) {

291 chomp;

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

293 my $name=$4;

294 my $comment=$6;

295 $dict{$name}=$comment;

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

297 }

298 else {

299 print F "$_\n";

300 }

301 }

302 }

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

305 @lines=<NOTES>;

306 close(NOTES);

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

309 for (@lines) {

310 chomp;

311 s/^\s+//;

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

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

314 print NOTES "$_\n";

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

316 }

317 }

318 for (keys %dict) {

319 $mark=$dict{$_};

320 $space=" "x(30-length($_));

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

322 }

323 }

325 PERL_SCRIPT

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

327 export NOTES_FILE

328 perl $PERL_SCRIPT_TEMP_NAME "$1"

329 rm $PERL_SCRIPT_TEMP_NAME

332 mkdir -p $WORK_DIR

333 oldpwd="$PWD"

334 cd $WORK_DIR

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

336 then

337 text_from_url "$1"

338 elif [ "$#" != 0 ]

339 then

340 if echo $1 | grep -q ^/

341 then

342 cat "$1"

343 else

344 cat "$oldpwd/$1"

345 fi

346 else

347 cat

348 fi \

349 | tee $ORIGINAL_TEXT \

350 | get_words ${TEMP1}-full \

351 | group_words \

352 | add_stat ${TEMP1}-full \

353 | tee "$TEMP1" > "$TEMP2"

355 if [ "$STAT_ONLY" = "YES" ]

356 then

357 cat "$TEMP1"

358 else

359 add_marks "$TEMP2"

360 if [ "$editor" = vim ]

361 then

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

363 else

364 echo 2

365 $editor "$TEMP2"

366 fi

367 remove_marks "$TEMP2"

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

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