1 // Copyright 2011 Google Inc. All Rights Reserved.
3 // Licensed under the Apache License, Version 2.0 (the "License");
4 // you may not use this file except in compliance with the License.
5 // You may obtain a copy of the License at
7 // http://www.apache.org/licenses/LICENSE-2.0
9 // Unless required by applicable law or agreed to in writing, software
10 // distributed under the License is distributed on an "AS IS" BASIS,
11 // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
12 // See the License for the specific language governing permissions and
13 // limitations under the License.
15 package com.hughes.android.dictionary.parser.enwiktionary;
17 import java.io.BufferedInputStream;
18 import java.io.DataInputStream;
19 import java.io.EOFException;
21 import java.io.FileInputStream;
22 import java.io.IOException;
23 import java.util.ArrayList;
24 import java.util.Arrays;
25 import java.util.Collection;
26 import java.util.LinkedHashSet;
27 import java.util.List;
30 import java.util.logging.Logger;
31 import java.util.regex.Pattern;
33 import com.hughes.android.dictionary.engine.EntryTypeName;
34 import com.hughes.android.dictionary.engine.IndexBuilder;
35 import com.hughes.android.dictionary.engine.IndexedEntry;
36 import com.hughes.android.dictionary.engine.PairEntry;
37 import com.hughes.android.dictionary.engine.PairEntry.Pair;
38 import com.hughes.android.dictionary.parser.WikiTokenizer;
40 public class EnWiktionaryXmlParser {
42 private static final String TRANSLITERATION_FORMAT = " (tr. %s)";
44 static final Logger LOG = Logger.getLogger(EnWiktionaryXmlParser.class.getName());
46 // TODO: process {{ttbc}} lines
48 static final Pattern partOfSpeechHeader = Pattern.compile(
49 "Noun|Verb|Adjective|Adverb|Pronoun|Conjunction|Interjection|" +
50 "Preposition|Proper noun|Article|Prepositional phrase|Acronym|" +
51 "Abbreviation|Initialism|Contraction|Prefix|Suffix|Symbol|Letter|" +
52 "Ligature|Idiom|Phrase|\\{\\{acronym\\}\\}|\\{\\{initialism\\}\\}|" +
53 // These are @deprecated:
54 "Noun form|Verb form|Adjective form|Nominal phrase|Noun phrase|" +
55 "Verb phrase|Transitive verb|Intransitive verb|Reflexive verb|" +
56 // These are extras I found:
57 "Determiner|Numeral|Number|Cardinal number|Ordinal number|Proverb|" +
58 "Particle|Interjection|Pronominal adverb" +
59 "Han character|Hanzi|Hanja|Kanji|Katakana character|Syllable");
61 final IndexBuilder enIndexBuilder;
62 final IndexBuilder foreignIndexBuilder;
63 final Pattern langPattern;
64 final Pattern langCodePattern;
67 public EnWiktionaryXmlParser(final IndexBuilder enIndexBuilder, final IndexBuilder otherIndexBuilder, final Pattern langPattern, final Pattern langCodePattern, final boolean swap) {
68 this.enIndexBuilder = enIndexBuilder;
69 this.foreignIndexBuilder = otherIndexBuilder;
70 this.langPattern = langPattern;
71 this.langCodePattern = langCodePattern;
76 public void parse(final File file, final int pageLimit) throws IOException {
78 final DataInputStream dis = new DataInputStream(new BufferedInputStream(new FileInputStream(file)));
80 if (pageLimit >= 0 && pageCount >= pageLimit) {
86 title = dis.readUTF();
87 } catch (EOFException e) {
91 final String heading = dis.readUTF();
92 final int bytesLength = dis.readInt();
93 final byte[] bytes = new byte[bytesLength];
95 final String text = new String(bytes, "UTF8");
97 parseSection(title, heading, text);
100 if (pageCount % 1000 == 0) {
101 LOG.info("pageCount=" + pageCount);
106 private void parseSection(final String title, String heading, final String text) {
107 if (title.startsWith("Wiktionary:") ||
108 title.startsWith("Template:") ||
109 title.startsWith("Appendix:") ||
110 title.startsWith("Category:") ||
111 title.startsWith("Index:") ||
112 title.startsWith("MediaWiki:") ||
113 title.startsWith("TransWiki:") ||
114 title.startsWith("Citations:") ||
115 title.startsWith("Concordance:") ||
116 title.startsWith("Help:")) {
120 heading = heading.replaceAll("=", "").trim();
121 if (heading.equals("English")) {
122 doEnglishWord(title, text);
123 } else if (langPattern.matcher(heading).find()){
124 doForeignWord(heading, title, text);
129 // -------------------------------------------------------------------------
131 private void doEnglishWord(String title, String text) {
136 final WikiTokenizer wikiTokenizer = new WikiTokenizer(text);
137 while (wikiTokenizer.nextToken() != null) {
139 if (wikiTokenizer.isHeading()) {
140 final String headerName = wikiTokenizer.headingWikiText();
142 if (wikiTokenizer.headingDepth() <= posDepth) {
147 if (partOfSpeechHeader.matcher(headerName).matches()) {
148 posDepth = wikiTokenizer.headingDepth();
149 pos = wikiTokenizer.headingWikiText();
150 // TODO: if we're inside the POS section, we should handle the first title line...
152 } else if (headerName.equals("Translations")) {
154 LOG.warning("Translations without POS: " + title);
156 doTranslations(title, wikiTokenizer, pos);
157 } else if (headerName.equals("Pronunciation")) {
158 //doPronunciation(wikiLineReader);
160 } else if (wikiTokenizer.isFunction()) {
161 final String name = wikiTokenizer.functionName();
162 if (name.equals("head")) {
163 LOG.warning("{{head}} without POS: " + title);
170 private static Set<String> encodings = new LinkedHashSet<String>(Arrays.asList("zh-ts",
171 "sd-Arab", "ku-Arab", "Arab", "unicode", "Laoo", "ur-Arab", "Thai",
172 "fa-Arab", "Khmr", "zh-tsp", "Cyrl", "IPAchar", "ug-Arab", "ko-inline",
173 "Jpan", "Kore", "Hebr", "rfscript", "Beng", "Mong", "Knda", "Cyrs",
174 "yue-tsj", "Mlym", "Tfng", "Grek", "yue-yue-j"));
176 private void doTranslations(final String title, final WikiTokenizer wikiTokenizer, final String pos) {
177 if (title.equals("absolutely")) {
178 //System.out.println();
181 String topLevelLang = null;
183 boolean done = false;
184 while (wikiTokenizer.nextToken() != null) {
185 if (wikiTokenizer.isHeading()) {
186 wikiTokenizer.returnToLineStart();
193 // Check whether we care about this line:
195 if (wikiTokenizer.isFunction()) {
196 final String functionName = wikiTokenizer.functionName();
197 final List<String> positionArgs = wikiTokenizer.functionPositionArgs();
199 if (functionName.equals("trans-top")) {
201 if (wikiTokenizer.functionPositionArgs().size() >= 1) {
202 sense = positionArgs.get(0);
203 // TODO: could emphasize words in [[brackets]] inside sense.
204 sense = WikiTokenizer.toPlainText(sense);
205 //LOG.info("Sense: " + sense);
207 } else if (functionName.equals("trans-bottom")) {
209 } else if (functionName.equals("trans-mid")) {
210 } else if (functionName.equals("trans-see")) {
211 // TODO: would also be nice...
212 } else if (functionName.startsWith("picdic")) {
213 } else if (functionName.startsWith("checktrans")) {
215 } else if (functionName.startsWith("ttbc")) {
216 wikiTokenizer.nextLine();
217 // TODO: would be great to handle ttbc
218 // TODO: Check this: done = true;
220 LOG.warning("Unexpected translation wikifunction: " + wikiTokenizer.token() + ", title=" + title);
222 } else if (wikiTokenizer.isListItem()) {
223 final String line = wikiTokenizer.listItemWikiText();
224 // This line could produce an output...
226 if (line.contains("ich hoan dich gear")) {
227 //System.out.println();
230 // First strip the language and check whether it matches.
231 // And hold onto it for sub-lines.
232 final int colonIndex = line.indexOf(":");
233 if (colonIndex == -1) {
237 final String lang = trim(WikiTokenizer.toPlainText(line.substring(0, colonIndex)));
238 final boolean appendLang;
239 if (wikiTokenizer.listItemPrefix().length() == 1) {
241 final boolean thisFind = langPattern.matcher(lang).find();
245 appendLang = !langPattern.matcher(lang).matches();
246 } else if (topLevelLang == null) {
249 // Two-level -- the only way we won't append is if this second level matches exactly.
250 if (!langPattern.matcher(lang).matches() && !langPattern.matcher(topLevelLang).find()) {
253 appendLang = !langPattern.matcher(lang).matches();
256 String rest = line.substring(colonIndex + 1).trim();
257 if (rest.length() > 0) {
258 doTranslationLine(line, appendLang ? lang : null, title, pos, sense, rest);
261 } else if (wikiTokenizer.remainderStartsWith("''See''")) {
262 wikiTokenizer.nextLine();
263 LOG.fine("Skipping See line: " + wikiTokenizer.token());
264 } else if (wikiTokenizer.isWikiLink()) {
265 final String wikiLink = wikiTokenizer.wikiLinkText();
266 if (wikiLink.contains(":") && wikiLink.contains(title)) {
267 } else if (wikiLink.contains("Category:")) {
269 LOG.warning("Unexpected wikiLink: " + wikiTokenizer.token() + ", title=" + title);
271 } else if (wikiTokenizer.isNewline() || wikiTokenizer.isMarkup() || wikiTokenizer.isComment()) {
273 final String token = wikiTokenizer.token();
274 if (token.equals("----")) {
276 LOG.warning("Unexpected translation token: " + wikiTokenizer.token() + ", title=" + title);
283 private void doTranslationLine(final String line, final String lang, final String title, final String pos, final String sense, final String rest) {
284 // Good chance we'll actually file this one...
285 final PairEntry pairEntry = new PairEntry();
286 final IndexedEntry indexedEntry = new IndexedEntry(pairEntry);
288 final StringBuilder foreignText = new StringBuilder();
289 final WikiTokenizer wikiTokenizer = new WikiTokenizer(rest, false);
290 while (wikiTokenizer.nextToken() != null) {
292 if (wikiTokenizer.isPlainText()) {
293 final String plainText = wikiTokenizer.token();
294 foreignText.append(plainText);
295 foreignIndexBuilder.addEntryWithString(indexedEntry, plainText, EntryTypeName.WIKTIONARY_TRANSLATION_OTHER_TEXT);
297 } else if (wikiTokenizer.isWikiLink()) {
298 final String plainText = wikiTokenizer.wikiLinkText();
299 foreignText.append(plainText);
300 // TODO: should check for English before appending.
301 foreignIndexBuilder.addEntryWithString(indexedEntry, plainText, EntryTypeName.WIKTIONARY_TRANSLATION_WIKI_TEXT);
303 } else if (wikiTokenizer.isFunction()) {
304 final String functionName = wikiTokenizer.functionName();
305 final List<String> args = wikiTokenizer.functionPositionArgs();
306 final Map<String,String> namedArgs = wikiTokenizer.functionNamedArgs();
308 if (functionName.equals("t") || functionName.equals("t+") || functionName.equals("t-") || functionName.equals("tø") || functionName.equals("apdx-t")) {
309 } else if (functionName.equals("qualifier")) {
310 } else if (encodings.contains(functionName)) {
311 foreignText.append("").append(args.get(0));
312 foreignIndexBuilder.addEntryWithString(indexedEntry, args.get(0), EntryTypeName.WIKTIONARY_TRANSLATION_OTHER_TEXT);
313 } else if (isGender(functionName)) {
314 appendGender(foreignText, functionName, args);
315 } else if (functionName.equals("g")) {
316 foreignText.append("{g}");
317 } else if (functionName.equals("l")) {
318 // encodes text in various langs.
320 foreignText.append("").append(args.get(1));
321 foreignIndexBuilder.addEntryWithString(indexedEntry, args.get(1), EntryTypeName.WIKTIONARY_TRANSLATION_OTHER_TEXT);
322 // TODO: transliteration
323 } else if (functionName.equals("term")) {
324 // cross-reference to another dictionary
325 foreignText.append("").append(args.get(0));
326 foreignIndexBuilder.addEntryWithString(indexedEntry, args.get(0), EntryTypeName.WIKTIONARY_TRANSLATION_OTHER_TEXT);
327 // TODO: transliteration
328 } else if (functionName.equals("italbrac") || functionName.equals("gloss")) {
329 // TODO: put this text aside to use it.
330 foreignText.append("[").append(args.get(0)).append("]");
331 foreignIndexBuilder.addEntryWithString(indexedEntry, args.get(0), EntryTypeName.WIKTIONARY_TRANSLATION_OTHER_TEXT);
332 } else if (functionName.equals("ttbc")) {
333 LOG.warning("Unexpected {{ttbc}}");
334 } else if (functionName.equals("trreq")) {
335 } else if (functionName.equals("not used")) {
336 foreignText.append("(not used)");
337 } else if (functionName.equals("t-image")) {
338 // American sign language
341 namedArgs.keySet().removeAll(USELESS_WIKI_ARGS);
342 WikiTokenizer.appendFunction(foreignText.append("{{"), functionName, args, namedArgs).append("}}");
345 } else if (wikiTokenizer.isNewline()) {
346 } else if (wikiTokenizer.isComment()) {
347 } else if (wikiTokenizer.isMarkup()) {
349 LOG.warning("Bad translation token: " + wikiTokenizer.token());
351 } // while-token loop.
354 if (foreignText.length() == 0) {
355 LOG.warning("Empty foreignText: " + line);
360 foreignText.insert(0, String.format("(%s) ", lang));
363 StringBuilder englishText = new StringBuilder();
365 englishText.append(title);
367 englishText.append(" (").append(sense).append(")");
368 enIndexBuilder.addEntryWithString(indexedEntry, sense, EntryTypeName.WIKTIONARY_TRANSLATION_SENSE, EntryTypeName.WIKTIONARY_TRANSLATION_SENSE);
371 englishText.append(" (").append(pos.toLowerCase()).append(")");
373 enIndexBuilder.addEntryWithString(indexedEntry, title, EntryTypeName.WIKTIONARY_TITLE_SINGLE, EntryTypeName.WIKTIONARY_TITLE_MULTI);
375 final Pair pair = new Pair(trim(englishText.toString()), trim(foreignText.toString()), swap);
376 pairEntry.pairs.add(pair);
377 if (!pairsAdded.add(pair.toString())) {
378 LOG.warning("Duplicate pair: " + pair.toString());
380 if (pair.toString().equals("libero {m} :: free (adjective)")) {
381 System.out.println();
387 private void appendGender(final StringBuilder otherText,
388 final String functionName, final List<String> args) {
389 otherText.append("{");
390 otherText.append(functionName);
391 for (int i = 0; i < args.size(); ++i) {
392 otherText.append("|").append(args.get(i));
394 otherText.append("}");
398 private boolean isGender(final String functionName) {
399 return functionName.equals("m") || functionName.equals("f") || functionName.equals("n") || functionName.equals("p");
402 Set<String> pairsAdded = new LinkedHashSet<String>();
404 // -------------------------------------------------------------------------
406 private void doForeignWord(final String lang, final String title, final String text) {
407 final WikiTokenizer wikiTokenizer = new WikiTokenizer(text);
408 while (wikiTokenizer.nextToken() != null) {
409 if (wikiTokenizer.isHeading()) {
410 final String headingName = wikiTokenizer.headingWikiText();
411 if (headingName.equals("Translations")) {
412 LOG.warning("Translations not in English section: " + title);
413 } else if (headingName.equals("Pronunciation")) {
414 //doPronunciation(wikiLineReader);
415 } else if (partOfSpeechHeader.matcher(headingName).matches()) {
416 doForeignPartOfSpeech(lang, title, headingName, wikiTokenizer.headingDepth(), wikiTokenizer);
423 static final class ListSection {
424 final String firstPrefix;
425 final String firstLine;
426 final List<String> nextPrefixes = new ArrayList<String>();
427 final List<String> nextLines = new ArrayList<String>();
429 public ListSection(String firstPrefix, String firstLine) {
430 this.firstPrefix = firstPrefix;
431 this.firstLine = firstLine;
435 public String toString() {
436 return firstPrefix + firstLine + "{ " + nextPrefixes + "}";
441 int foreignCount = 0;
442 private void doForeignPartOfSpeech(final String lang, String title, final String posHeading, final int posDepth, WikiTokenizer wikiTokenizer) {
443 if (++foreignCount % 1000 == 0) {
444 LOG.info("***" + lang + ", " + title + ", pos=" + posHeading + ", foreignCount=" + foreignCount);
446 if (title.equals("moro")) {
447 System.out.println();
450 boolean titleAppended = false;
451 final StringBuilder foreignBuilder = new StringBuilder();
452 final Collection<String> wordForms = new ArrayList<String>();
453 final List<ListSection> listSections = new ArrayList<ListSection>();
457 ListSection lastListSection = null;
459 int currentHeadingDepth = posDepth;
460 while (wikiTokenizer.nextToken() != null) {
461 if (wikiTokenizer.isHeading()) {
462 currentHeadingDepth = wikiTokenizer.headingDepth();
464 if (currentHeadingDepth <= posDepth) {
465 wikiTokenizer.returnToLineStart();
470 if (currentHeadingDepth > posDepth) {
471 // TODO: deal with other neat info sections
475 if (wikiTokenizer.isFunction()) {
476 final String name = wikiTokenizer.functionName();
477 final List<String> args = wikiTokenizer.functionPositionArgs();
478 final Map<String,String> namedArgs = wikiTokenizer.functionNamedArgs();
479 // First line is generally a repeat of the title with some extra information.
480 // We need to build up the left side (foreign text, tokens) separately from the
481 // right side (English). The left-side may get paired with multiple right sides.
482 // The left side should get filed under every form of the word in question (singular, plural).
484 // For verbs, the conjugation comes later on in a deeper section.
485 // Ideally, we'd want to file every English entry with the verb
486 // under every verb form coming from the conjugation.
487 // Ie. under "fa": see: "make :: fare" and "do :: fare"
488 // But then where should we put the conjugation table?
489 // I think just under fare. But then we need a way to link to the entry (actually the row, since entries doesn't show up!)
490 // for the conjugation table from "fa".
491 // Would like to be able to link to a lang#token.
492 if (isGender(name)) {
493 appendGender(foreignBuilder, name, args);
494 } else if (name.equals("wikipedia")) {
495 namedArgs.remove("lang");
496 if (args.size() > 1 || !namedArgs.isEmpty()) {
498 foreignBuilder.append(wikiTokenizer.token());
499 } else if (args.size() == 1) {
500 foreignBuilder.append(wikiTokenizer.token());
502 //foreignBuilder.append(title);
504 } else if (name.equals("attention") || name.equals("zh-attention")) {
505 // See: http://en.wiktionary.org/wiki/Template:attention
507 } else if (name.equals("infl") || name.equals("head")) {
508 // See: http://en.wiktionary.org/wiki/Template:infl
509 final String langCode = get(args, 0);
510 String head = namedArgs.remove("head");
512 head = namedArgs.remove("title"); // Bug
517 head = WikiTokenizer.toPlainText(head);
519 titleAppended = true;
521 namedArgs.keySet().removeAll(USELESS_WIKI_ARGS);
523 final String tr = namedArgs.remove("tr");
524 String g = namedArgs.remove("g");
526 g = namedArgs.remove("gender");
528 final String g2 = namedArgs.remove("g2");
529 final String g3 = namedArgs.remove("g3");
531 foreignBuilder.append(head);
534 foreignBuilder.append(" {").append(g);
536 foreignBuilder.append("|").append(g2);
539 foreignBuilder.append("|").append(g3);
541 foreignBuilder.append("}");
545 foreignBuilder.append(String.format(TRANSLITERATION_FORMAT, tr));
549 final String pos = get(args, 1);
551 foreignBuilder.append(" (").append(pos).append(")");
553 for (int i = 2; i < args.size(); i += 2) {
554 final String inflName = get(args, i);
555 final String inflValue = get(args, i + 1);
556 foreignBuilder.append(", ").append(WikiTokenizer.toPlainText(inflName));
557 if (inflValue != null && inflValue.length() > 0) {
558 foreignBuilder.append(": ").append(WikiTokenizer.toPlainText(inflValue));
559 wordForms.add(inflValue);
562 for (final String key : namedArgs.keySet()) {
563 final String value = WikiTokenizer.toPlainText(namedArgs.get(key));
564 foreignBuilder.append(" ").append(key).append("=").append(value);
565 wordForms.add(value);
567 } else if (name.equals("it-noun")) {
568 titleAppended = true;
569 final String base = get(args, 0);
570 final String gender = get(args, 1);
571 final String singular = base + get(args, 2);
572 final String plural = base + get(args, 3);
573 foreignBuilder.append(String.format(" %s {%s}, %s {pl}", singular, gender, plural, plural));
574 wordForms.add(singular);
575 wordForms.add(plural);
576 if (!namedArgs.isEmpty() || args.size() > 4) {
577 LOG.warning("Invalid it-noun: " + wikiTokenizer.token());
579 } else if (name.equals("it-proper noun")) {
580 foreignBuilder.append(wikiTokenizer.token());
581 } else if (name.equals("it-adj")) {
582 foreignBuilder.append(wikiTokenizer.token());
583 } else if (name.startsWith("it-conj")) {
584 if (name.equals("it-conj-are")) {
585 itConjAre(args, namedArgs);
586 } else if (name.equals("it-conj-ere")) {
587 } else if (name.equals("it-conj-ire")) {
589 LOG.warning("Unknown conjugation: " + wikiTokenizer.token());
593 foreignBuilder.append(wikiTokenizer.token());
594 // LOG.warning("Unknown function: " + wikiTokenizer.token());
596 } else if (wikiTokenizer.isListItem()) {
597 final String prefix = wikiTokenizer.listItemPrefix();
598 if (lastListSection != null &&
599 prefix.startsWith(lastListSection.firstPrefix) &&
600 prefix.length() > lastListSection.firstPrefix.length()) {
601 lastListSection.nextPrefixes.add(prefix);
602 lastListSection.nextLines.add(wikiTokenizer.listItemWikiText());
604 lastListSection = new ListSection(prefix, wikiTokenizer.listItemWikiText());
605 listSections.add(lastListSection);
607 } else if (lastListSection != null) {
608 // Don't append anything after the lists, because there's crap.
609 } else if (wikiTokenizer.isWikiLink()) {
611 foreignBuilder.append(wikiTokenizer.wikiLinkText());
613 } else if (wikiTokenizer.isPlainText()) {
615 foreignBuilder.append(wikiTokenizer.token());
617 } else if (wikiTokenizer.isMarkup() || wikiTokenizer.isNewline() || wikiTokenizer.isComment()) {
620 LOG.warning("Unexpected token: " + wikiTokenizer.token());
625 // Here's where we exit.
626 // Should we make an entry even if there are no foreign list items?
627 String foreign = foreignBuilder.toString().trim();
628 if (!titleAppended && !foreign.toLowerCase().startsWith(title.toLowerCase())) {
629 foreign = String.format("%s %s", title, foreign);
631 if (!langPattern.matcher(lang).matches()) {
632 foreign = String.format("(%s) %s", lang, foreign);
634 for (final ListSection listSection : listSections) {
635 doForeignListItem(foreign, title, wordForms, listSection);
641 static final Pattern UNINDEXED_WIKI_TEXT = Pattern.compile(
642 "(first|second|third)-person (singular|plural)|" +
647 // Might only want to remove "lang" if it's equal to "zh", for example.
648 static final Set<String> USELESS_WIKI_ARGS = new LinkedHashSet<String>(Arrays.asList("lang", "sc", "sort", "cat"));
650 private void doForeignListItem(final String foreignText, String title, final Collection<String> forms, final ListSection listSection) {
652 final String prefix = listSection.firstPrefix;
653 if (prefix.length() > 1) {
654 // Could just get looser and say that any prefix longer than first is a sublist.
655 LOG.warning("Prefix too long: " + listSection);
659 final PairEntry pairEntry = new PairEntry();
660 final IndexedEntry indexedEntry = new IndexedEntry(pairEntry);
662 final StringBuilder englishBuilder = new StringBuilder();
664 final String mainLine = listSection.firstLine;
665 final WikiTokenizer englishTokenizer = new WikiTokenizer(mainLine, false);
666 while (englishTokenizer.nextToken() != null) {
667 // TODO handle form of....
668 if (englishTokenizer.isPlainText()) {
669 englishBuilder.append(englishTokenizer.token());
670 enIndexBuilder.addEntryWithString(indexedEntry, englishTokenizer.token(), EntryTypeName.WIKTIONARY_ENGLISH_DEF);
671 } else if (englishTokenizer.isWikiLink()) {
672 final String text = englishTokenizer.wikiLinkText();
673 final String link = englishTokenizer.wikiLinkDest();
675 if (link.contains("#English")) {
676 englishBuilder.append(text);
677 enIndexBuilder.addEntryWithString(indexedEntry, text, EntryTypeName.WIKTIONARY_ENGLISH_DEF_WIKI_LINK);
678 } else if (link.contains("#") && this.langPattern.matcher(link).find()) {
679 englishBuilder.append(text);
680 foreignIndexBuilder.addEntryWithString(indexedEntry, text, EntryTypeName.WIKTIONARY_ENGLISH_DEF_OTHER_LANG);
681 } else if (link.equals("plural")) {
682 englishBuilder.append(text);
684 //LOG.warning("Special link: " + englishTokenizer.token());
685 enIndexBuilder.addEntryWithString(indexedEntry, text, EntryTypeName.WIKTIONARY_ENGLISH_DEF_WIKI_LINK);
686 englishBuilder.append(text);
690 englishBuilder.append(text);
691 if (!UNINDEXED_WIKI_TEXT.matcher(text).find()) {
692 enIndexBuilder.addEntryWithString(indexedEntry, text, EntryTypeName.WIKTIONARY_ENGLISH_DEF_WIKI_LINK);
695 } else if (englishTokenizer.isFunction()) {
696 final String name = englishTokenizer.functionName();
697 final List<String> args = englishTokenizer.functionPositionArgs();
698 final Map<String,String> namedArgs = englishTokenizer.functionNamedArgs();
701 name.equals("form of") ||
702 name.contains("conjugation of") ||
703 name.contains("participle of") ||
704 name.contains("gerund of") ||
705 name.contains("feminine of") ||
706 name.contains("plural of")) {
707 String formName = name;
708 if (name.equals("form of")) {
709 formName = remove(args, 0, null);
711 if (formName == null) {
712 LOG.warning("Missing form name: " + title);
713 formName = "form of";
715 String baseForm = get(args, 1, "");
716 if ("".equals(baseForm)) {
717 baseForm = get(args, 0, null);
720 remove(args, 0, null);
722 namedArgs.keySet().removeAll(USELESS_WIKI_ARGS);
723 WikiTokenizer.appendFunction(englishBuilder.append("{"), formName, args, namedArgs).append("}");
724 if (baseForm != null) {
725 foreignIndexBuilder.addEntryWithString(indexedEntry, baseForm, EntryTypeName.WIKTIONARY_BASE_FORM_SINGLE, EntryTypeName.WIKTIONARY_BASE_FORM_MULTI);
727 // null baseForm happens in Danish.
728 LOG.warning("Null baseform: " + title);
730 } else if (name.equals("l")) {
731 // encodes text in various langs.
733 englishBuilder.append("").append(args.get(1));
734 final String langCode = args.get(0);
735 if ("en".equals(langCode)) {
736 enIndexBuilder.addEntryWithString(indexedEntry, args.get(1), EntryTypeName.WIKTIONARY_ENGLISH_DEF_WIKI_LINK);
738 foreignIndexBuilder.addEntryWithString(indexedEntry, args.get(1), EntryTypeName.WIKTIONARY_ENGLISH_DEF_OTHER_LANG);
740 // TODO: transliteration
742 } else if (name.equals("defn") || name.equals("rfdef")) {
744 // http://en.wiktionary.org/wiki/Wiktionary:Requests_for_deletion/Others#Template:defn
745 // Redundant, used for the same purpose as {{rfdef}}, but this
746 // doesn't produce the "This word needs a definition" text.
747 // Delete or redirect.
749 namedArgs.keySet().removeAll(USELESS_WIKI_ARGS);
750 if (args.size() == 0 && namedArgs.isEmpty()) {
751 englishBuilder.append("{").append(name).append("}");
753 WikiTokenizer.appendFunction(englishBuilder.append("{{"), name, args, namedArgs).append("}}");
755 // LOG.warning("Unexpected function: " + englishTokenizer.token());
758 if (englishTokenizer.isComment() || englishTokenizer.isMarkup()) {
760 LOG.warning("Unexpected definition type: " + englishTokenizer.token());
765 final String english = trim(englishBuilder.toString());
766 if (english.length() > 0) {
767 final Pair pair = new Pair(english, trim(foreignText), this.swap);
768 pairEntry.pairs.add(pair);
769 foreignIndexBuilder.addEntryWithString(indexedEntry, title, EntryTypeName.WIKTIONARY_TITLE_SINGLE, EntryTypeName.WIKTIONARY_TITLE_MULTI);
770 for (final String form : forms) {
771 foreignIndexBuilder.addEntryWithString(indexedEntry, form, EntryTypeName.WIKTIONARY_INFLECTD_FORM_SINGLE, EntryTypeName.WIKTIONARY_INFLECTED_FORM_MULTI);
776 String lastForeign = null;
777 for (int i = 0; i < listSection.nextPrefixes.size(); ++i) {
778 final String nextPrefix = listSection.nextPrefixes.get(i);
779 final String nextLine = listSection.nextLines.get(i);
780 int dash = nextLine.indexOf("—");
783 dash = nextLine.indexOf("—");
787 dash = nextLine.indexOf(" - ");
791 if ((nextPrefix.equals("#:") || nextPrefix.equals("##:")) && dash != -1) {
792 final String foreignEx = nextLine.substring(0, dash);
793 final String englishEx = nextLine.substring(dash + mdashLen);
794 final Pair pair = new Pair(formatAndIndexExampleString(englishEx, enIndexBuilder, indexedEntry), formatAndIndexExampleString(foreignEx, foreignIndexBuilder, indexedEntry), swap);
795 if (pair.lang1 != "--" && pair.lang1 != "--") {
796 pairEntry.pairs.add(pair);
799 } else if (nextPrefix.equals("#:") || nextPrefix.equals("##:")){
800 final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
801 lastForeign = nextLine;
802 if (pair.lang1 != "--" && pair.lang1 != "--") {
803 pairEntry.pairs.add(pair);
805 } else if (nextPrefix.equals("#::") || nextPrefix.equals("#**")) {
806 if (lastForeign != null && pairEntry.pairs.size() > 0) {
807 pairEntry.pairs.remove(pairEntry.pairs.size() - 1);
808 final Pair pair = new Pair(formatAndIndexExampleString(nextLine, enIndexBuilder, indexedEntry), formatAndIndexExampleString(lastForeign, foreignIndexBuilder, indexedEntry), swap);
809 if (pair.lang1 != "--" || pair.lang2 != "--") {
810 pairEntry.pairs.add(pair);
814 LOG.warning("TODO: English example with no foreign: " + title + ", " + nextLine);
815 final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
816 if (pair.lang1 != "--" || pair.lang2 != "--") {
817 pairEntry.pairs.add(pair);
820 } else if (nextPrefix.equals("#*")) {
821 // Can't really index these.
822 final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
823 lastForeign = nextLine;
824 if (pair.lang1 != "--" || pair.lang2 != "--") {
825 pairEntry.pairs.add(pair);
827 } else if (nextPrefix.equals("#::*") || nextPrefix.equals("##") || nextPrefix.equals("#*:") || nextPrefix.equals("#:*") || true) {
828 final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
829 if (pair.lang1 != "--" || pair.lang2 != "--") {
830 pairEntry.pairs.add(pair);
838 private String formatAndIndexExampleString(final String example, final IndexBuilder indexBuilder, final IndexedEntry indexedEntry) {
839 final WikiTokenizer wikiTokenizer = new WikiTokenizer(example, false);
840 final StringBuilder builder = new StringBuilder();
841 boolean insideTripleQuotes = false;
842 while (wikiTokenizer.nextToken() != null) {
843 if (wikiTokenizer.isPlainText()) {
844 builder.append(wikiTokenizer.token());
845 if (indexBuilder != null) {
846 indexBuilder.addEntryWithString(indexedEntry, wikiTokenizer.token(), EntryTypeName.WIKTIONARY_EXAMPLE);
848 } else if (wikiTokenizer.isWikiLink()) {
849 final String text = wikiTokenizer.wikiLinkText().replaceAll("'", "");
850 builder.append(text);
851 if (indexBuilder != null) {
852 indexBuilder.addEntryWithString(indexedEntry, text, EntryTypeName.WIKTIONARY_EXAMPLE);
854 } else if (wikiTokenizer.isFunction()) {
855 builder.append(wikiTokenizer.token());
856 } else if (wikiTokenizer.isMarkup()) {
857 if (wikiTokenizer.token().equals("'''")) {
858 insideTripleQuotes = !insideTripleQuotes;
860 } else if (wikiTokenizer.isComment() || wikiTokenizer.isNewline()) {
863 LOG.warning("unexpected token: " + wikiTokenizer.token());
866 final String result = trim(builder.toString());
867 return result.length() > 0 ? result : "--";
871 private void itConjAre(List<String> args, Map<String, String> namedArgs) {
872 final String base = args.get(0);
873 final String aux = args.get(1);
875 putIfMissing(namedArgs, "inf", base + "are");
876 putIfMissing(namedArgs, "aux", aux);
877 putIfMissing(namedArgs, "ger", base + "ando");
878 putIfMissing(namedArgs, "presp", base + "ante");
879 putIfMissing(namedArgs, "pastp", base + "ato");
881 putIfMissing(namedArgs, "pres1s", base + "o");
882 putIfMissing(namedArgs, "pres2s", base + "i");
883 putIfMissing(namedArgs, "pres3s", base + "a");
884 putIfMissing(namedArgs, "pres1p", base + "iamo");
885 putIfMissing(namedArgs, "pres2p", base + "ate");
886 putIfMissing(namedArgs, "pres3p", base + "ano");
888 putIfMissing(namedArgs, "imperf1s", base + "avo");
889 putIfMissing(namedArgs, "imperf2s", base + "avi");
890 putIfMissing(namedArgs, "imperf3s", base + "ava");
891 putIfMissing(namedArgs, "imperf1p", base + "avamo");
892 putIfMissing(namedArgs, "imperf2p", base + "avate");
893 putIfMissing(namedArgs, "imperf3p", base + "avano");
895 putIfMissing(namedArgs, "prem1s", base + "ai");
896 putIfMissing(namedArgs, "prem2s", base + "asti");
897 putIfMissing(namedArgs, "prem3s", base + "ò");
898 putIfMissing(namedArgs, "prem1p", base + "ammo");
899 putIfMissing(namedArgs, "prem2p", base + "aste");
900 putIfMissing(namedArgs, "prem3p", base + "arono");
902 putIfMissing(namedArgs, "fut1s", base + "erò");
903 putIfMissing(namedArgs, "fut2s", base + "erai");
904 putIfMissing(namedArgs, "fut3s", base + "erà");
905 putIfMissing(namedArgs, "fut1p", base + "eremo");
906 putIfMissing(namedArgs, "fut2p", base + "erete");
907 putIfMissing(namedArgs, "fut3p", base + "eranno");
909 putIfMissing(namedArgs, "cond1s", base + "erei");
910 putIfMissing(namedArgs, "cond2s", base + "eresti");
911 putIfMissing(namedArgs, "cond3s", base + "erebbe");
912 putIfMissing(namedArgs, "cond1p", base + "eremmo");
913 putIfMissing(namedArgs, "cond2p", base + "ereste");
914 putIfMissing(namedArgs, "cond3p", base + "erebbero");
915 // Subjunctive / congiuntivo
916 putIfMissing(namedArgs, "sub123s", base + "i");
917 putIfMissing(namedArgs, "sub1p", base + "iamo");
918 putIfMissing(namedArgs, "sub2p", base + "iate");
919 putIfMissing(namedArgs, "sub3p", base + "ino");
920 // Imperfect subjunctive
921 putIfMissing(namedArgs, "impsub12s", base + "assi");
922 putIfMissing(namedArgs, "impsub3s", base + "asse");
923 putIfMissing(namedArgs, "impsub1p", base + "assimo");
924 putIfMissing(namedArgs, "impsub2p", base + "aste");
925 putIfMissing(namedArgs, "impsub3p", base + "assero");
927 putIfMissing(namedArgs, "imp2s", base + "a");
928 putIfMissing(namedArgs, "imp3s", base + "i");
929 putIfMissing(namedArgs, "imp1p", base + "iamo");
930 putIfMissing(namedArgs, "imp2p", base + "ate");
931 putIfMissing(namedArgs, "imp3p", base + "ino");
934 itConj(args, namedArgs);
938 private void itConj(List<String> args, Map<String, String> namedArgs) {
939 // TODO Auto-generated method stub
944 private static void putIfMissing(final Map<String, String> namedArgs, final String key,
945 final String value) {
946 final String oldValue = namedArgs.get(key);
947 if (oldValue == null || oldValue.length() == 0) {
948 namedArgs.put(key, value);
952 // TODO: check how ='' and =| are manifested....
953 // TODO: get this right in -are
954 private static void putOrNullify(final Map<String, String> namedArgs, final String key,
955 final String value) {
956 final String oldValue = namedArgs.get(key);
957 if (oldValue == null/* || oldValue.length() == 0*/) {
958 namedArgs.put(key, value);
960 if (oldValue.equals("''")) {
961 namedArgs.put(key, "");
966 static final Pattern whitespace = Pattern.compile("\\s+");
967 static String trim(final String s) {
968 return whitespace.matcher(s).replaceAll(" ").trim();