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;
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;
39 public class EnWiktionaryXmlParser {
41 private static final String TRANSLITERATION_FORMAT = " (tr. %s)";
43 static final Logger LOG = Logger.getLogger(EnWiktionaryXmlParser.class.getName());
45 // TODO: process {{ttbc}} lines
47 static final Pattern partOfSpeechHeader = Pattern.compile(
48 "Noun|Verb|Adjective|Adverb|Pronoun|Conjunction|Interjection|" +
49 "Preposition|Proper noun|Article|Prepositional phrase|Acronym|" +
50 "Abbreviation|Initialism|Contraction|Prefix|Suffix|Symbol|Letter|" +
51 "Ligature|Idiom|Phrase|\\{\\{acronym\\}\\}|\\{\\{initialism\\}\\}|" +
52 // These are @deprecated:
53 "Noun form|Verb form|Adjective form|Nominal phrase|Noun phrase|" +
54 "Verb phrase|Transitive verb|Intransitive verb|Reflexive verb|" +
55 // These are extras I found:
56 "Determiner|Numeral|Number|Cardinal number|Ordinal number|Proverb|" +
57 "Particle|Interjection|Pronominal adverb" +
58 "Han character|Hanzi|Hanja|Kanji|Katakana character|Syllable");
60 final IndexBuilder enIndexBuilder;
61 final IndexBuilder foreignIndexBuilder;
62 final Pattern langPattern;
63 final Pattern langCodePattern;
66 public EnWiktionaryXmlParser(final IndexBuilder enIndexBuilder, final IndexBuilder otherIndexBuilder, final Pattern langPattern, final Pattern langCodePattern, final boolean swap) {
67 this.enIndexBuilder = enIndexBuilder;
68 this.foreignIndexBuilder = otherIndexBuilder;
69 this.langPattern = langPattern;
70 this.langCodePattern = langCodePattern;
75 public void parse(final File file, final int pageLimit) throws IOException {
77 final DataInputStream dis = new DataInputStream(new BufferedInputStream(new FileInputStream(file)));
79 if (pageLimit >= 0 && pageCount >= pageLimit) {
85 title = dis.readUTF();
86 } catch (EOFException e) {
90 final String heading = dis.readUTF();
91 final int bytesLength = dis.readInt();
92 final byte[] bytes = new byte[bytesLength];
94 final String text = new String(bytes, "UTF8");
96 parseSection(title, heading, text);
99 if (pageCount % 1000 == 0) {
100 LOG.info("pageCount=" + pageCount);
105 private void parseSection(final String title, String heading, final String text) {
106 if (title.startsWith("Wiktionary:") ||
107 title.startsWith("Template:") ||
108 title.startsWith("Appendix:") ||
109 title.startsWith("Category:") ||
110 title.startsWith("Index:") ||
111 title.startsWith("MediaWiki:") ||
112 title.startsWith("TransWiki:") ||
113 title.startsWith("Citations:") ||
114 title.startsWith("Concordance:") ||
115 title.startsWith("Help:")) {
119 heading = heading.replaceAll("=", "").trim();
120 if (heading.equals("English")) {
121 doEnglishWord(title, text);
122 } else if (langPattern.matcher(heading).find()){
123 doForeignWord(heading, title, text);
128 // -------------------------------------------------------------------------
130 private void doEnglishWord(String title, String text) {
135 final WikiTokenizer wikiTokenizer = new WikiTokenizer(text);
136 while (wikiTokenizer.nextToken() != null) {
138 if (wikiTokenizer.isHeading()) {
139 final String headerName = wikiTokenizer.headingWikiText();
141 if (wikiTokenizer.headingDepth() <= posDepth) {
146 if (partOfSpeechHeader.matcher(headerName).matches()) {
147 posDepth = wikiTokenizer.headingDepth();
148 pos = wikiTokenizer.headingWikiText();
149 // TODO: if we're inside the POS section, we should handle the first title line...
151 } else if (headerName.equals("Translations")) {
153 LOG.warning("Translations without POS: " + title);
155 doTranslations(title, wikiTokenizer, pos);
156 } else if (headerName.equals("Pronunciation")) {
157 //doPronunciation(wikiLineReader);
164 private static Set<String> encodings = new LinkedHashSet<String>(Arrays.asList("zh-ts",
165 "sd-Arab", "ku-Arab", "Arab", "unicode", "Laoo", "ur-Arab", "Thai",
166 "fa-Arab", "Khmr", "zh-tsp", "Cyrl", "IPAchar", "ug-Arab", "ko-inline",
167 "Jpan", "Kore", "Hebr", "rfscript", "Beng", "Mong", "Knda", "Cyrs",
168 "yue-tsj", "Mlym", "Tfng", "Grek", "yue-yue-j"));
170 private void doTranslations(final String title, final WikiTokenizer wikiTokenizer, final String pos) {
171 if (title.equals("absolutely")) {
172 //System.out.println();
175 String topLevelLang = null;
177 boolean done = false;
178 while (wikiTokenizer.nextToken() != null) {
179 if (wikiTokenizer.isHeading()) {
180 wikiTokenizer.returnToLineStart();
187 // Check whether we care about this line:
189 if (wikiTokenizer.isFunction()) {
190 final String functionName = wikiTokenizer.functionName();
191 final List<String> positionArgs = wikiTokenizer.functionPositionArgs();
193 if (functionName.equals("trans-top")) {
195 if (wikiTokenizer.functionPositionArgs().size() >= 1) {
196 sense = positionArgs.get(0);
197 // TODO: could emphasize words in [[brackets]] inside sense.
198 sense = WikiTokenizer.toPlainText(sense);
199 //LOG.info("Sense: " + sense);
201 } else if (functionName.equals("trans-bottom")) {
203 } else if (functionName.equals("trans-mid")) {
204 } else if (functionName.equals("trans-see")) {
205 // TODO: would also be nice...
206 } else if (functionName.startsWith("picdic")) {
207 } else if (functionName.startsWith("checktrans")) {
209 } else if (functionName.startsWith("ttbc")) {
210 wikiTokenizer.nextLine();
211 // TODO: would be great to handle ttbc
212 // TODO: Check this: done = true;
214 LOG.warning("Unexpected translation wikifunction: " + wikiTokenizer.token() + ", title=" + title);
216 } else if (wikiTokenizer.isListItem()) {
217 final String line = wikiTokenizer.listItemWikiText();
218 // This line could produce an output...
220 if (line.contains("ich hoan dich gear")) {
221 //System.out.println();
224 // First strip the language and check whether it matches.
225 // And hold onto it for sub-lines.
226 final int colonIndex = line.indexOf(":");
227 if (colonIndex == -1) {
231 final String lang = trim(WikiTokenizer.toPlainText(line.substring(0, colonIndex)));
232 final boolean appendLang;
233 if (wikiTokenizer.listItemPrefix().length() == 1) {
235 final boolean thisFind = langPattern.matcher(lang).find();
239 appendLang = !langPattern.matcher(lang).matches();
240 } else if (topLevelLang == null) {
243 // Two-level -- the only way we won't append is if this second level matches exactly.
244 if (!langPattern.matcher(lang).matches() && !langPattern.matcher(topLevelLang).find()) {
247 appendLang = !langPattern.matcher(lang).matches();
250 String rest = line.substring(colonIndex + 1).trim();
251 if (rest.length() > 0) {
252 doTranslationLine(line, appendLang ? lang : null, title, pos, sense, rest);
255 } else if (wikiTokenizer.remainderStartsWith("''See''")) {
256 wikiTokenizer.nextLine();
257 LOG.fine("Skipping line: " + wikiTokenizer.token());
258 } else if (wikiTokenizer.isWikiLink()) {
259 final String wikiLink = wikiTokenizer.wikiLinkText();
260 if (wikiLink.contains(":") && wikiLink.contains(title)) {
261 } else if (wikiLink.contains("Category:")) {
263 LOG.warning("Unexpected wikiLink: " + wikiTokenizer.token() + ", title=" + title);
265 } else if (wikiTokenizer.isNewline() || wikiTokenizer.isMarkup() || wikiTokenizer.isComment()) {
267 final String token = wikiTokenizer.token();
268 if (token.equals("----")) {
270 LOG.warning("Unexpected translation token: " + wikiTokenizer.token() + ", title=" + title);
277 private static <T> T get(final List<T> list, final int index, final T defaultValue) {
278 return index < list.size() ? list.get(index) : defaultValue;
281 private static <T> T get(final List<T> list, final int index) {
282 return get(list, index, null);
285 private static <T> T remove(final List<T> list, final int index, final T defaultValue) {
286 return index < list.size() ? list.remove(index) : defaultValue;
290 static final class Callback implements WikiTokenizer.Callback {
291 public Callback(IndexedEntry indexedEntry, IndexBuilder defaultIndexBuilder,
292 StringBuilder builder, Map<String, WikiFunctionCallback> functionCallbacks) {
293 this.indexedEntry = indexedEntry;
294 this.defaultIndexBuilder = defaultIndexBuilder;
295 this.builder = builder;
296 this.functionCallbacks = functionCallbacks;
299 final IndexedEntry indexedEntry;
300 final IndexBuilder defaultIndexBuilder;
301 final StringBuilder builder;
302 final Map<String,WikiFunctionCallback> functionCallbacks;
304 // TODO: the classes of text are wrong....
307 public void onPlainText(WikiTokenizer wikiTokenizer) {
308 final String plainText = wikiTokenizer.token();
309 builder.append(plainText);
310 defaultIndexBuilder.addEntryWithString(indexedEntry, plainText, EntryTypeName.WIKTIONARY_TRANSLATION_OTHER_TEXT);
314 public void onWikiLink(WikiTokenizer wikiTokenizer) {
315 final String plainText = wikiTokenizer.wikiLinkText();
316 builder.append(plainText);
317 // TODO: should check for English before appending.
318 defaultIndexBuilder.addEntryWithString(indexedEntry, plainText, EntryTypeName.WIKTIONARY_TRANSLATION_WIKI_TEXT);
322 public void onFunction(String functionName,
323 List<String> functionPositionArgs, Map<String, String> functionNamedArgs) {
327 public void onMarkup(WikiTokenizer wikiTokenizer) {
332 public void onComment(WikiTokenizer wikiTokenizer) {
337 public void onNewline(WikiTokenizer wikiTokenizer) {
342 public void onHeading(WikiTokenizer wikiTokenizer) {
347 public void onListItem(WikiTokenizer wikiTokenizer) {
352 private void doTranslationLine(final String line, final String lang, final String title, final String pos, final String sense, final String rest) {
353 // Good chance we'll actually file this one...
354 final PairEntry pairEntry = new PairEntry();
355 final IndexedEntry indexedEntry = new IndexedEntry(pairEntry);
357 final StringBuilder foreignText = new StringBuilder();
358 final WikiTokenizer wikiTokenizer = new WikiTokenizer(rest, false);
359 while (wikiTokenizer.nextToken() != null) {
361 if (wikiTokenizer.isPlainText()) {
362 final String plainText = wikiTokenizer.token();
363 foreignText.append(plainText);
364 foreignIndexBuilder.addEntryWithString(indexedEntry, plainText, EntryTypeName.WIKTIONARY_TRANSLATION_OTHER_TEXT);
366 } else if (wikiTokenizer.isWikiLink()) {
367 final String plainText = wikiTokenizer.wikiLinkText();
368 foreignText.append(plainText);
369 // TODO: should check for English before appending.
370 foreignIndexBuilder.addEntryWithString(indexedEntry, plainText, EntryTypeName.WIKTIONARY_TRANSLATION_WIKI_TEXT);
372 } else if (wikiTokenizer.isFunction()) {
373 final String functionName = wikiTokenizer.functionName();
374 final List<String> args = wikiTokenizer.functionPositionArgs();
375 final Map<String,String> namedArgs = wikiTokenizer.functionNamedArgs();
377 if (functionName.equals("t") || functionName.equals("t+") || functionName.equals("t-") || functionName.equals("tø") || functionName.equals("apdx-t")) {
378 if (args.size() < 2) {
379 LOG.warning("{{t}} with too few args: " + line + ", title=" + title);
382 final String langCode = get(args, 0);
383 final String word = get(args, 1);
384 final String gender = get(args, 2);
385 final String transliteration = namedArgs.get("tr");
386 if (foreignText.length() > 0) {
387 foreignText.append("");
389 foreignText.append(word);
390 foreignIndexBuilder.addEntryWithString(indexedEntry, word, EntryTypeName.WIKTIONARY_TITLE_SINGLE, EntryTypeName.WIKTIONARY_TITLE_MULTI);
391 if (gender != null) {
392 foreignText.append(String.format(" {%s}", gender));
394 if (transliteration != null) {
395 foreignText.append(String.format(TRANSLITERATION_FORMAT, transliteration));
396 foreignIndexBuilder.addEntryWithString(indexedEntry, transliteration, EntryTypeName.WIKTIONARY_TRANSLITERATION);
398 } else if (functionName.equals("qualifier")) {
399 if (args.size() == 0) {
400 foreignText.append(wikiTokenizer.token());
402 String qualifier = args.get(0);
403 if (!namedArgs.isEmpty() || args.size() > 1) {
404 LOG.warning("weird qualifier: " + line);
407 foreignText.append("(").append(qualifier).append(")");
409 } else if (encodings.contains(functionName)) {
410 foreignText.append("").append(args.get(0));
411 foreignIndexBuilder.addEntryWithString(indexedEntry, args.get(0), EntryTypeName.WIKTIONARY_TRANSLATION_OTHER_TEXT);
412 } else if (isGender(functionName)) {
413 appendGender(foreignText, functionName, args);
414 } else if (functionName.equals("g")) {
415 foreignText.append("{g}");
416 } else if (functionName.equals("l")) {
417 // encodes text in various langs.
419 foreignText.append("").append(args.get(1));
420 foreignIndexBuilder.addEntryWithString(indexedEntry, args.get(1), EntryTypeName.WIKTIONARY_TRANSLATION_OTHER_TEXT);
421 // TODO: transliteration
422 } else if (functionName.equals("term")) {
423 // cross-reference to another dictionary
424 foreignText.append("").append(args.get(0));
425 foreignIndexBuilder.addEntryWithString(indexedEntry, args.get(0), EntryTypeName.WIKTIONARY_TRANSLATION_OTHER_TEXT);
426 // TODO: transliteration
427 } else if (functionName.equals("italbrac") || functionName.equals("gloss")) {
428 // TODO: put this text aside to use it.
429 foreignText.append("[").append(args.get(0)).append("]");
430 foreignIndexBuilder.addEntryWithString(indexedEntry, args.get(0), EntryTypeName.WIKTIONARY_TRANSLATION_OTHER_TEXT);
431 } else if (functionName.equals("ttbc")) {
432 LOG.warning("Unexpected {{ttbc}}");
433 } else if (functionName.equals("trreq")) {
434 } else if (functionName.equals("not used")) {
435 foreignText.append("(not used)");
436 } else if (functionName.equals("t-image")) {
437 // American sign language
440 namedArgs.keySet().removeAll(USELESS_WIKI_ARGS);
441 WikiTokenizer.appendFunction(foreignText.append("{{"), functionName, args, namedArgs).append("}}");
444 } else if (wikiTokenizer.isNewline()) {
445 } else if (wikiTokenizer.isComment()) {
446 } else if (wikiTokenizer.isMarkup()) {
448 LOG.warning("Bad translation token: " + wikiTokenizer.token());
451 if (foreignText.length() == 0) {
452 LOG.warning("Empty foreignText: " + line);
457 foreignText.insert(0, String.format("(%s) ", lang));
460 StringBuilder englishText = new StringBuilder();
462 englishText.append(title);
464 englishText.append(" (").append(sense).append(")");
465 enIndexBuilder.addEntryWithString(indexedEntry, sense, EntryTypeName.WIKTIONARY_TRANSLATION_SENSE, EntryTypeName.WIKTIONARY_TRANSLATION_SENSE);
468 englishText.append(" (").append(pos.toLowerCase()).append(")");
470 enIndexBuilder.addEntryWithString(indexedEntry, title, EntryTypeName.WIKTIONARY_TITLE_SINGLE, EntryTypeName.WIKTIONARY_TITLE_MULTI);
472 final Pair pair = new Pair(trim(englishText.toString()), trim(foreignText.toString()), swap);
473 pairEntry.pairs.add(pair);
474 if (!pairsAdded.add(pair.toString())) {
475 LOG.warning("Duplicate pair: " + pair.toString());
477 if (pair.toString().equals("libero {m} :: free (adjective)")) {
478 System.out.println();
484 private void appendGender(final StringBuilder otherText,
485 final String functionName, final List<String> args) {
486 otherText.append("{");
487 otherText.append(functionName);
488 for (int i = 0; i < args.size(); ++i) {
489 otherText.append("|").append(args.get(i));
491 otherText.append("}");
495 private boolean isGender(final String functionName) {
496 return functionName.equals("m") || functionName.equals("f") || functionName.equals("n") || functionName.equals("p");
499 Set<String> pairsAdded = new LinkedHashSet<String>();
501 // -------------------------------------------------------------------------
503 private void doForeignWord(final String lang, final String title, final String text) {
504 final WikiTokenizer wikiTokenizer = new WikiTokenizer(text);
505 while (wikiTokenizer.nextToken() != null) {
506 if (wikiTokenizer.isHeading()) {
507 final String headingName = wikiTokenizer.headingWikiText();
508 if (headingName.equals("Translations")) {
509 LOG.warning("Translations not in English section: " + title);
510 } else if (headingName.equals("Pronunciation")) {
511 //doPronunciation(wikiLineReader);
512 } else if (partOfSpeechHeader.matcher(headingName).matches()) {
513 doForeignPartOfSpeech(lang, title, headingName, wikiTokenizer.headingDepth(), wikiTokenizer);
520 static final class ListSection {
521 final String firstPrefix;
522 final String firstLine;
523 final List<String> nextPrefixes = new ArrayList<String>();
524 final List<String> nextLines = new ArrayList<String>();
526 public ListSection(String firstPrefix, String firstLine) {
527 this.firstPrefix = firstPrefix;
528 this.firstLine = firstLine;
532 public String toString() {
533 return firstPrefix + firstLine + "{ " + nextPrefixes + "}";
538 int foreignCount = 0;
539 private void doForeignPartOfSpeech(final String lang, String title, final String posHeading, final int posDepth, WikiTokenizer wikiTokenizer) {
540 if (++foreignCount % 1000 == 0) {
541 LOG.info("***" + lang + ", " + title + ", pos=" + posHeading + ", foreignCount=" + foreignCount);
543 if (title.equals("moro")) {
544 System.out.println();
547 boolean titleAppended = false;
548 final StringBuilder foreignBuilder = new StringBuilder();
549 final Collection<String> wordForms = new ArrayList<String>();
550 final List<ListSection> listSections = new ArrayList<ListSection>();
554 ListSection lastListSection = null;
556 int currentHeadingDepth = posDepth;
557 while (wikiTokenizer.nextToken() != null) {
558 if (wikiTokenizer.isHeading()) {
559 currentHeadingDepth = wikiTokenizer.headingDepth();
561 if (currentHeadingDepth <= posDepth) {
562 wikiTokenizer.returnToLineStart();
567 if (currentHeadingDepth > posDepth) {
568 // TODO: deal with other neat info sections
572 if (wikiTokenizer.isFunction()) {
573 final String name = wikiTokenizer.functionName();
574 final List<String> args = wikiTokenizer.functionPositionArgs();
575 final Map<String,String> namedArgs = wikiTokenizer.functionNamedArgs();
576 // First line is generally a repeat of the title with some extra information.
577 // We need to build up the left side (foreign text, tokens) separately from the
578 // right side (English). The left-side may get paired with multiple right sides.
579 // The left side should get filed under every form of the word in question (singular, plural).
581 // For verbs, the conjugation comes later on in a deeper section.
582 // Ideally, we'd want to file every English entry with the verb
583 // under every verb form coming from the conjugation.
584 // Ie. under "fa": see: "make :: fare" and "do :: fare"
585 // But then where should we put the conjugation table?
586 // 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!)
587 // for the conjugation table from "fa".
588 // Would like to be able to link to a lang#token.
589 if (isGender(name)) {
590 appendGender(foreignBuilder, name, args);
591 } else if (name.equals("wikipedia")) {
592 namedArgs.remove("lang");
593 if (args.size() > 1 || !namedArgs.isEmpty()) {
595 foreignBuilder.append(wikiTokenizer.token());
596 } else if (args.size() == 1) {
597 foreignBuilder.append(wikiTokenizer.token());
599 //foreignBuilder.append(title);
601 } else if (name.equals("attention") || name.equals("zh-attention")) {
602 // See: http://en.wiktionary.org/wiki/Template:attention
604 } else if (name.equals("infl") || name.equals("head")) {
605 // See: http://en.wiktionary.org/wiki/Template:infl
606 final String langCode = get(args, 0);
607 String head = namedArgs.remove("head");
609 head = namedArgs.remove("title"); // Bug
614 head = WikiTokenizer.toPlainText(head);
616 titleAppended = true;
618 namedArgs.keySet().removeAll(USELESS_WIKI_ARGS);
620 final String tr = namedArgs.remove("tr");
621 String g = namedArgs.remove("g");
623 g = namedArgs.remove("gender");
625 final String g2 = namedArgs.remove("g2");
626 final String g3 = namedArgs.remove("g3");
628 foreignBuilder.append(head);
631 foreignBuilder.append(" {").append(g);
633 foreignBuilder.append("|").append(g2);
636 foreignBuilder.append("|").append(g3);
638 foreignBuilder.append("}");
642 foreignBuilder.append(String.format(TRANSLITERATION_FORMAT, tr));
646 final String pos = get(args, 1);
648 foreignBuilder.append(" (").append(pos).append(")");
650 for (int i = 2; i < args.size(); i += 2) {
651 final String inflName = get(args, i);
652 final String inflValue = get(args, i + 1);
653 foreignBuilder.append(", ").append(WikiTokenizer.toPlainText(inflName));
654 if (inflValue != null && inflValue.length() > 0) {
655 foreignBuilder.append(": ").append(WikiTokenizer.toPlainText(inflValue));
656 wordForms.add(inflValue);
659 for (final String key : namedArgs.keySet()) {
660 final String value = WikiTokenizer.toPlainText(namedArgs.get(key));
661 foreignBuilder.append(" ").append(key).append("=").append(value);
662 wordForms.add(value);
664 } else if (name.equals("it-noun")) {
665 titleAppended = true;
666 final String base = get(args, 0);
667 final String gender = get(args, 1);
668 final String singular = base + get(args, 2);
669 final String plural = base + get(args, 3);
670 foreignBuilder.append(String.format(" %s {%s}, %s {pl}", singular, gender, plural, plural));
671 wordForms.add(singular);
672 wordForms.add(plural);
673 if (!namedArgs.isEmpty() || args.size() > 4) {
674 LOG.warning("Invalid it-noun: " + wikiTokenizer.token());
676 } else if (name.equals("it-proper noun")) {
677 foreignBuilder.append(wikiTokenizer.token());
678 } else if (name.equals("it-adj")) {
679 foreignBuilder.append(wikiTokenizer.token());
680 } else if (name.startsWith("it-conj")) {
681 if (name.equals("it-conj-are")) {
682 itConjAre(args, namedArgs);
683 } else if (name.equals("it-conj-ere")) {
684 } else if (name.equals("it-conj-ire")) {
686 LOG.warning("Unknown conjugation: " + wikiTokenizer.token());
690 foreignBuilder.append(wikiTokenizer.token());
691 // LOG.warning("Unknown function: " + wikiTokenizer.token());
693 } else if (wikiTokenizer.isListItem()) {
694 final String prefix = wikiTokenizer.listItemPrefix();
695 if (lastListSection != null &&
696 prefix.startsWith(lastListSection.firstPrefix) &&
697 prefix.length() > lastListSection.firstPrefix.length()) {
698 lastListSection.nextPrefixes.add(prefix);
699 lastListSection.nextLines.add(wikiTokenizer.listItemWikiText());
701 lastListSection = new ListSection(prefix, wikiTokenizer.listItemWikiText());
702 listSections.add(lastListSection);
704 } else if (lastListSection != null) {
705 // Don't append anything after the lists, because there's crap.
706 } else if (wikiTokenizer.isWikiLink()) {
708 foreignBuilder.append(wikiTokenizer.wikiLinkText());
710 } else if (wikiTokenizer.isPlainText()) {
712 foreignBuilder.append(wikiTokenizer.token());
714 } else if (wikiTokenizer.isMarkup() || wikiTokenizer.isNewline() || wikiTokenizer.isComment()) {
717 LOG.warning("Unexpected token: " + wikiTokenizer.token());
722 // Here's where we exit.
723 // Should we make an entry even if there are no foreign list items?
724 String foreign = foreignBuilder.toString().trim();
725 if (!titleAppended && !foreign.toLowerCase().startsWith(title.toLowerCase())) {
726 foreign = String.format("%s %s", title, foreign);
728 if (!langPattern.matcher(lang).matches()) {
729 foreign = String.format("(%s) %s", lang, foreign);
731 for (final ListSection listSection : listSections) {
732 doForeignListItem(foreign, title, wordForms, listSection);
738 static final Pattern UNINDEXED_WIKI_TEXT = Pattern.compile(
739 "(first|second|third)-person (singular|plural)|" +
744 // Might only want to remove "lang" if it's equal to "zh", for example.
745 static final Set<String> USELESS_WIKI_ARGS = new LinkedHashSet<String>(Arrays.asList("lang", "sc", "sort", "cat"));
747 private void doForeignListItem(final String foreignText, String title, final Collection<String> forms, final ListSection listSection) {
749 final String prefix = listSection.firstPrefix;
750 if (prefix.length() > 1) {
751 // Could just get looser and say that any prefix longer than first is a sublist.
752 LOG.warning("Prefix too long: " + listSection);
756 final PairEntry pairEntry = new PairEntry();
757 final IndexedEntry indexedEntry = new IndexedEntry(pairEntry);
759 final StringBuilder englishBuilder = new StringBuilder();
761 final String mainLine = listSection.firstLine;
762 final WikiTokenizer englishTokenizer = new WikiTokenizer(mainLine, false);
763 while (englishTokenizer.nextToken() != null) {
764 // TODO handle form of....
765 if (englishTokenizer.isPlainText()) {
766 englishBuilder.append(englishTokenizer.token());
767 enIndexBuilder.addEntryWithString(indexedEntry, englishTokenizer.token(), EntryTypeName.WIKTIONARY_ENGLISH_DEF);
768 } else if (englishTokenizer.isWikiLink()) {
769 final String text = englishTokenizer.wikiLinkText();
770 final String link = englishTokenizer.wikiLinkDest();
772 if (link.contains("#English")) {
773 englishBuilder.append(text);
774 enIndexBuilder.addEntryWithString(indexedEntry, text, EntryTypeName.WIKTIONARY_ENGLISH_DEF_WIKI_LINK);
775 } else if (link.contains("#") && this.langPattern.matcher(link).find()) {
776 englishBuilder.append(text);
777 foreignIndexBuilder.addEntryWithString(indexedEntry, text, EntryTypeName.WIKTIONARY_ENGLISH_DEF_OTHER_LANG);
778 } else if (link.equals("plural")) {
779 englishBuilder.append(text);
781 //LOG.warning("Special link: " + englishTokenizer.token());
782 enIndexBuilder.addEntryWithString(indexedEntry, text, EntryTypeName.WIKTIONARY_ENGLISH_DEF_WIKI_LINK);
783 englishBuilder.append(text);
787 englishBuilder.append(text);
788 if (!UNINDEXED_WIKI_TEXT.matcher(text).find()) {
789 enIndexBuilder.addEntryWithString(indexedEntry, text, EntryTypeName.WIKTIONARY_ENGLISH_DEF_WIKI_LINK);
792 } else if (englishTokenizer.isFunction()) {
793 final String name = englishTokenizer.functionName();
794 final List<String> args = englishTokenizer.functionPositionArgs();
795 final Map<String,String> namedArgs = englishTokenizer.functionNamedArgs();
798 name.equals("form of") ||
799 name.contains("conjugation of") ||
800 name.contains("participle of") ||
801 name.contains("gerund of") ||
802 name.contains("feminine of") ||
803 name.contains("plural of")) {
804 String formName = name;
805 if (name.equals("form of")) {
806 formName = remove(args, 0, null);
808 if (formName == null) {
809 LOG.warning("Missing form name: " + title);
810 formName = "form of";
812 String baseForm = get(args, 1, "");
813 if ("".equals(baseForm)) {
814 baseForm = get(args, 0, null);
817 remove(args, 0, null);
819 namedArgs.keySet().removeAll(USELESS_WIKI_ARGS);
820 WikiTokenizer.appendFunction(englishBuilder.append("{"), formName, args, namedArgs).append("}");
821 if (baseForm != null) {
822 foreignIndexBuilder.addEntryWithString(indexedEntry, baseForm, EntryTypeName.WIKTIONARY_BASE_FORM_SINGLE, EntryTypeName.WIKTIONARY_BASE_FORM_MULTI);
824 // null baseForm happens in Danish.
825 LOG.warning("Null baseform: " + title);
827 } else if (name.equals("l")) {
828 // encodes text in various langs.
830 englishBuilder.append("").append(args.get(1));
831 final String langCode = args.get(0);
832 if ("en".equals(langCode)) {
833 enIndexBuilder.addEntryWithString(indexedEntry, args.get(1), EntryTypeName.WIKTIONARY_ENGLISH_DEF_WIKI_LINK);
835 foreignIndexBuilder.addEntryWithString(indexedEntry, args.get(1), EntryTypeName.WIKTIONARY_ENGLISH_DEF_OTHER_LANG);
837 // TODO: transliteration
839 } else if (name.equals("defn") || name.equals("rfdef")) {
841 // http://en.wiktionary.org/wiki/Wiktionary:Requests_for_deletion/Others#Template:defn
842 // Redundant, used for the same purpose as {{rfdef}}, but this
843 // doesn't produce the "This word needs a definition" text.
844 // Delete or redirect.
846 namedArgs.keySet().removeAll(USELESS_WIKI_ARGS);
847 if (args.size() == 0 && namedArgs.isEmpty()) {
848 englishBuilder.append("{").append(name).append("}");
850 WikiTokenizer.appendFunction(englishBuilder.append("{{"), name, args, namedArgs).append("}}");
852 // LOG.warning("Unexpected function: " + englishTokenizer.token());
855 if (englishTokenizer.isComment() || englishTokenizer.isMarkup()) {
857 LOG.warning("Unexpected definition type: " + englishTokenizer.token());
862 final String english = trim(englishBuilder.toString());
863 if (english.length() > 0) {
864 final Pair pair = new Pair(english, trim(foreignText), this.swap);
865 pairEntry.pairs.add(pair);
866 foreignIndexBuilder.addEntryWithString(indexedEntry, title, EntryTypeName.WIKTIONARY_TITLE_SINGLE, EntryTypeName.WIKTIONARY_TITLE_MULTI);
867 for (final String form : forms) {
868 foreignIndexBuilder.addEntryWithString(indexedEntry, form, EntryTypeName.WIKTIONARY_INFLECTD_FORM_SINGLE, EntryTypeName.WIKTIONARY_INFLECTED_FORM_MULTI);
873 String lastForeign = null;
874 for (int i = 0; i < listSection.nextPrefixes.size(); ++i) {
875 final String nextPrefix = listSection.nextPrefixes.get(i);
876 final String nextLine = listSection.nextLines.get(i);
877 int dash = nextLine.indexOf("—");
880 dash = nextLine.indexOf("—");
884 dash = nextLine.indexOf(" - ");
888 if ((nextPrefix.equals("#:") || nextPrefix.equals("##:")) && dash != -1) {
889 final String foreignEx = nextLine.substring(0, dash);
890 final String englishEx = nextLine.substring(dash + mdashLen);
891 final Pair pair = new Pair(formatAndIndexExampleString(englishEx, enIndexBuilder, indexedEntry), formatAndIndexExampleString(foreignEx, foreignIndexBuilder, indexedEntry), swap);
892 if (pair.lang1 != "--" && pair.lang1 != "--") {
893 pairEntry.pairs.add(pair);
896 } else if (nextPrefix.equals("#:") || nextPrefix.equals("##:")){
897 final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
898 lastForeign = nextLine;
899 if (pair.lang1 != "--" && pair.lang1 != "--") {
900 pairEntry.pairs.add(pair);
902 } else if (nextPrefix.equals("#::") || nextPrefix.equals("#**")) {
903 if (lastForeign != null && pairEntry.pairs.size() > 0) {
904 pairEntry.pairs.remove(pairEntry.pairs.size() - 1);
905 final Pair pair = new Pair(formatAndIndexExampleString(nextLine, enIndexBuilder, indexedEntry), formatAndIndexExampleString(lastForeign, foreignIndexBuilder, indexedEntry), swap);
906 if (pair.lang1 != "--" || pair.lang2 != "--") {
907 pairEntry.pairs.add(pair);
911 LOG.warning("TODO: English example with no foreign: " + title + ", " + nextLine);
912 final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
913 if (pair.lang1 != "--" || pair.lang2 != "--") {
914 pairEntry.pairs.add(pair);
917 } else if (nextPrefix.equals("#*")) {
918 // Can't really index these.
919 final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
920 lastForeign = nextLine;
921 if (pair.lang1 != "--" || pair.lang2 != "--") {
922 pairEntry.pairs.add(pair);
924 } else if (nextPrefix.equals("#::*") || nextPrefix.equals("##") || nextPrefix.equals("#*:") || nextPrefix.equals("#:*") || true) {
925 final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
926 if (pair.lang1 != "--" || pair.lang2 != "--") {
927 pairEntry.pairs.add(pair);
935 private String formatAndIndexExampleString(final String example, final IndexBuilder indexBuilder, final IndexedEntry indexedEntry) {
936 final WikiTokenizer wikiTokenizer = new WikiTokenizer(example, false);
937 final StringBuilder builder = new StringBuilder();
938 boolean insideTripleQuotes = false;
939 while (wikiTokenizer.nextToken() != null) {
940 if (wikiTokenizer.isPlainText()) {
941 builder.append(wikiTokenizer.token());
942 if (indexBuilder != null) {
943 indexBuilder.addEntryWithString(indexedEntry, wikiTokenizer.token(), EntryTypeName.WIKTIONARY_EXAMPLE);
945 } else if (wikiTokenizer.isWikiLink()) {
946 final String text = wikiTokenizer.wikiLinkText().replaceAll("'", "");
947 builder.append(text);
948 if (indexBuilder != null) {
949 indexBuilder.addEntryWithString(indexedEntry, text, EntryTypeName.WIKTIONARY_EXAMPLE);
951 } else if (wikiTokenizer.isFunction()) {
952 builder.append(wikiTokenizer.token());
953 } else if (wikiTokenizer.isMarkup()) {
954 if (wikiTokenizer.token().equals("'''")) {
955 insideTripleQuotes = !insideTripleQuotes;
957 } else if (wikiTokenizer.isComment() || wikiTokenizer.isNewline()) {
960 LOG.warning("unexpected token: " + wikiTokenizer.token());
963 final String result = trim(builder.toString());
964 return result.length() > 0 ? result : "--";
968 private void itConjAre(List<String> args, Map<String, String> namedArgs) {
969 final String base = args.get(0);
970 final String aux = args.get(1);
972 putIfMissing(namedArgs, "inf", base + "are");
973 putIfMissing(namedArgs, "aux", aux);
974 putIfMissing(namedArgs, "ger", base + "ando");
975 putIfMissing(namedArgs, "presp", base + "ante");
976 putIfMissing(namedArgs, "pastp", base + "ato");
978 putIfMissing(namedArgs, "pres1s", base + "o");
979 putIfMissing(namedArgs, "pres2s", base + "i");
980 putIfMissing(namedArgs, "pres3s", base + "a");
981 putIfMissing(namedArgs, "pres1p", base + "iamo");
982 putIfMissing(namedArgs, "pres2p", base + "ate");
983 putIfMissing(namedArgs, "pres3p", base + "ano");
985 putIfMissing(namedArgs, "imperf1s", base + "avo");
986 putIfMissing(namedArgs, "imperf2s", base + "avi");
987 putIfMissing(namedArgs, "imperf3s", base + "ava");
988 putIfMissing(namedArgs, "imperf1p", base + "avamo");
989 putIfMissing(namedArgs, "imperf2p", base + "avate");
990 putIfMissing(namedArgs, "imperf3p", base + "avano");
992 putIfMissing(namedArgs, "prem1s", base + "ai");
993 putIfMissing(namedArgs, "prem2s", base + "asti");
994 putIfMissing(namedArgs, "prem3s", base + "ò");
995 putIfMissing(namedArgs, "prem1p", base + "ammo");
996 putIfMissing(namedArgs, "prem2p", base + "aste");
997 putIfMissing(namedArgs, "prem3p", base + "arono");
999 putIfMissing(namedArgs, "fut1s", base + "erò");
1000 putIfMissing(namedArgs, "fut2s", base + "erai");
1001 putIfMissing(namedArgs, "fut3s", base + "erà");
1002 putIfMissing(namedArgs, "fut1p", base + "eremo");
1003 putIfMissing(namedArgs, "fut2p", base + "erete");
1004 putIfMissing(namedArgs, "fut3p", base + "eranno");
1006 putIfMissing(namedArgs, "cond1s", base + "erei");
1007 putIfMissing(namedArgs, "cond2s", base + "eresti");
1008 putIfMissing(namedArgs, "cond3s", base + "erebbe");
1009 putIfMissing(namedArgs, "cond1p", base + "eremmo");
1010 putIfMissing(namedArgs, "cond2p", base + "ereste");
1011 putIfMissing(namedArgs, "cond3p", base + "erebbero");
1012 // Subjunctive / congiuntivo
1013 putIfMissing(namedArgs, "sub123s", base + "i");
1014 putIfMissing(namedArgs, "sub1p", base + "iamo");
1015 putIfMissing(namedArgs, "sub2p", base + "iate");
1016 putIfMissing(namedArgs, "sub3p", base + "ino");
1017 // Imperfect subjunctive
1018 putIfMissing(namedArgs, "impsub12s", base + "assi");
1019 putIfMissing(namedArgs, "impsub3s", base + "asse");
1020 putIfMissing(namedArgs, "impsub1p", base + "assimo");
1021 putIfMissing(namedArgs, "impsub2p", base + "aste");
1022 putIfMissing(namedArgs, "impsub3p", base + "assero");
1024 putIfMissing(namedArgs, "imp2s", base + "a");
1025 putIfMissing(namedArgs, "imp3s", base + "i");
1026 putIfMissing(namedArgs, "imp1p", base + "iamo");
1027 putIfMissing(namedArgs, "imp2p", base + "ate");
1028 putIfMissing(namedArgs, "imp3p", base + "ino");
1031 itConj(args, namedArgs);
1035 private void itConj(List<String> args, Map<String, String> namedArgs) {
1036 // TODO Auto-generated method stub
1041 private static void putIfMissing(final Map<String, String> namedArgs, final String key,
1042 final String value) {
1043 final String oldValue = namedArgs.get(key);
1044 if (oldValue == null || oldValue.length() == 0) {
1045 namedArgs.put(key, value);
1049 // TODO: check how ='' and =| are manifested....
1050 // TODO: get this right in -are
1051 private static void putOrNullify(final Map<String, String> namedArgs, final String key,
1052 final String value) {
1053 final String oldValue = namedArgs.get(key);
1054 if (oldValue == null/* || oldValue.length() == 0*/) {
1055 namedArgs.put(key, value);
1057 if (oldValue.equals("''")) {
1058 namedArgs.put(key, "");
1063 static final Pattern whitespace = Pattern.compile("\\s+");
1064 static String trim(final String s) {
1065 return whitespace.matcher(s).replaceAll(" ").trim();