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 final Map<String,WikiFunctionCallback> functionCallbacks;
292 final StringBuilder builder;
293 final IndexBuilder defaultIndexBuilder;
294 final IndexedEntry indexedEntry;
296 // TODO: the classes of text are wrong....
299 public void onPlainText(WikiTokenizer wikiTokenizer) {
300 final String plainText = wikiTokenizer.token();
301 builder.append(plainText);
302 defaultIndexBuilder.addEntryWithString(indexedEntry, plainText, EntryTypeName.WIKTIONARY_TRANSLATION_OTHER_TEXT);
306 public void onWikiLink(WikiTokenizer wikiTokenizer) {
307 final String plainText = wikiTokenizer.wikiLinkText();
308 builder.append(plainText);
309 // TODO: should check for English before appending.
310 defaultIndexBuilder.addEntryWithString(indexedEntry, plainText, EntryTypeName.WIKTIONARY_TRANSLATION_WIKI_TEXT);
314 public void onFunction(String functionName,
315 List<String> functionPositionArgs, Map<String, String> functionNamedArgs) {
319 public void onMarkup(WikiTokenizer wikiTokenizer) {
324 public void onComment(WikiTokenizer wikiTokenizer) {
329 public void onNewline(WikiTokenizer wikiTokenizer) {
334 public void onHeading(WikiTokenizer wikiTokenizer) {
339 public void onListItem(WikiTokenizer wikiTokenizer) {
344 private void doTranslationLine(final String line, final String lang, final String title, final String pos, final String sense, final String rest) {
345 // Good chance we'll actually file this one...
346 final PairEntry pairEntry = new PairEntry();
347 final IndexedEntry indexedEntry = new IndexedEntry(pairEntry);
349 final StringBuilder foreignText = new StringBuilder();
350 final WikiTokenizer wikiTokenizer = new WikiTokenizer(rest, false);
351 while (wikiTokenizer.nextToken() != null) {
353 if (wikiTokenizer.isPlainText()) {
355 } else if (wikiTokenizer.isWikiLink()) {
357 } else if (wikiTokenizer.isFunction()) {
358 final String functionName = wikiTokenizer.functionName();
359 final List<String> args = wikiTokenizer.functionPositionArgs();
360 final Map<String,String> namedArgs = wikiTokenizer.functionNamedArgs();
362 if (functionName.equals("t") || functionName.equals("t+") || functionName.equals("t-") || functionName.equals("tø") || functionName.equals("apdx-t")) {
363 if (args.size() < 2) {
364 LOG.warning("{{t}} with too few args: " + line + ", title=" + title);
367 final String langCode = get(args, 0);
368 final String word = get(args, 1);
369 final String gender = get(args, 2);
370 final String transliteration = namedArgs.get("tr");
371 if (foreignText.length() > 0) {
372 foreignText.append("");
374 foreignText.append(word);
375 foreignIndexBuilder.addEntryWithString(indexedEntry, word, EntryTypeName.WIKTIONARY_TITLE_SINGLE, EntryTypeName.WIKTIONARY_TITLE_MULTI);
376 if (gender != null) {
377 foreignText.append(String.format(" {%s}", gender));
379 if (transliteration != null) {
380 foreignText.append(String.format(TRANSLITERATION_FORMAT, transliteration));
381 foreignIndexBuilder.addEntryWithString(indexedEntry, transliteration, EntryTypeName.WIKTIONARY_TRANSLITERATION);
383 } else if (functionName.equals("qualifier")) {
384 if (args.size() == 0) {
385 foreignText.append(wikiTokenizer.token());
387 String qualifier = args.get(0);
388 if (!namedArgs.isEmpty() || args.size() > 1) {
389 LOG.warning("weird qualifier: " + line);
392 foreignText.append("(").append(qualifier).append(")");
394 } else if (encodings.contains(functionName)) {
395 foreignText.append("").append(args.get(0));
396 foreignIndexBuilder.addEntryWithString(indexedEntry, args.get(0), EntryTypeName.WIKTIONARY_TRANSLATION_OTHER_TEXT);
397 } else if (isGender(functionName)) {
398 appendGender(foreignText, functionName, args);
399 } else if (functionName.equals("g")) {
400 foreignText.append("{g}");
401 } else if (functionName.equals("l")) {
402 // encodes text in various langs.
404 foreignText.append("").append(args.get(1));
405 foreignIndexBuilder.addEntryWithString(indexedEntry, args.get(1), EntryTypeName.WIKTIONARY_TRANSLATION_OTHER_TEXT);
406 // TODO: transliteration
407 } else if (functionName.equals("term")) {
408 // cross-reference to another dictionary
409 foreignText.append("").append(args.get(0));
410 foreignIndexBuilder.addEntryWithString(indexedEntry, args.get(0), EntryTypeName.WIKTIONARY_TRANSLATION_OTHER_TEXT);
411 // TODO: transliteration
412 } else if (functionName.equals("italbrac") || functionName.equals("gloss")) {
413 // TODO: put this text aside to use it.
414 foreignText.append("[").append(args.get(0)).append("]");
415 foreignIndexBuilder.addEntryWithString(indexedEntry, args.get(0), EntryTypeName.WIKTIONARY_TRANSLATION_OTHER_TEXT);
416 } else if (functionName.equals("ttbc")) {
417 LOG.warning("Unexpected {{ttbc}}");
418 } else if (functionName.equals("trreq")) {
419 } else if (functionName.equals("not used")) {
420 foreignText.append("(not used)");
421 } else if (functionName.equals("t-image")) {
422 // American sign language
425 namedArgs.keySet().removeAll(USELESS_WIKI_ARGS);
426 WikiTokenizer.appendFunction(foreignText.append("{{"), functionName, args, namedArgs).append("}}");
429 } else if (wikiTokenizer.isNewline()) {
430 } else if (wikiTokenizer.isComment()) {
431 } else if (wikiTokenizer.isMarkup()) {
433 LOG.warning("Bad translation token: " + wikiTokenizer.token());
436 if (foreignText.length() == 0) {
437 LOG.warning("Empty foreignText: " + line);
442 foreignText.insert(0, String.format("(%s) ", lang));
445 StringBuilder englishText = new StringBuilder();
447 englishText.append(title);
449 englishText.append(" (").append(sense).append(")");
450 enIndexBuilder.addEntryWithString(indexedEntry, sense, EntryTypeName.WIKTIONARY_TRANSLATION_SENSE, EntryTypeName.WIKTIONARY_TRANSLATION_SENSE);
453 englishText.append(" (").append(pos.toLowerCase()).append(")");
455 enIndexBuilder.addEntryWithString(indexedEntry, title, EntryTypeName.WIKTIONARY_TITLE_SINGLE, EntryTypeName.WIKTIONARY_TITLE_MULTI);
457 final Pair pair = new Pair(trim(englishText.toString()), trim(foreignText.toString()), swap);
458 pairEntry.pairs.add(pair);
459 if (!pairsAdded.add(pair.toString())) {
460 LOG.warning("Duplicate pair: " + pair.toString());
462 if (pair.toString().equals("libero {m} :: free (adjective)")) {
463 System.out.println();
469 private void appendGender(final StringBuilder otherText,
470 final String functionName, final List<String> args) {
471 otherText.append("{");
472 otherText.append(functionName);
473 for (int i = 0; i < args.size(); ++i) {
474 otherText.append("|").append(args.get(i));
476 otherText.append("}");
480 private boolean isGender(final String functionName) {
481 return functionName.equals("m") || functionName.equals("f") || functionName.equals("n") || functionName.equals("p");
484 Set<String> pairsAdded = new LinkedHashSet<String>();
486 // -------------------------------------------------------------------------
488 private void doForeignWord(final String lang, final String title, final String text) {
489 final WikiTokenizer wikiTokenizer = new WikiTokenizer(text);
490 while (wikiTokenizer.nextToken() != null) {
491 if (wikiTokenizer.isHeading()) {
492 final String headingName = wikiTokenizer.headingWikiText();
493 if (headingName.equals("Translations")) {
494 LOG.warning("Translations not in English section: " + title);
495 } else if (headingName.equals("Pronunciation")) {
496 //doPronunciation(wikiLineReader);
497 } else if (partOfSpeechHeader.matcher(headingName).matches()) {
498 doForeignPartOfSpeech(lang, title, headingName, wikiTokenizer.headingDepth(), wikiTokenizer);
505 static final class ListSection {
506 final String firstPrefix;
507 final String firstLine;
508 final List<String> nextPrefixes = new ArrayList<String>();
509 final List<String> nextLines = new ArrayList<String>();
511 public ListSection(String firstPrefix, String firstLine) {
512 this.firstPrefix = firstPrefix;
513 this.firstLine = firstLine;
517 public String toString() {
518 return firstPrefix + firstLine + "{ " + nextPrefixes + "}";
523 int foreignCount = 0;
524 private void doForeignPartOfSpeech(final String lang, String title, final String posHeading, final int posDepth, WikiTokenizer wikiTokenizer) {
525 if (++foreignCount % 1000 == 0) {
526 LOG.info("***" + lang + ", " + title + ", pos=" + posHeading + ", foreignCount=" + foreignCount);
528 if (title.equals("moro")) {
529 System.out.println();
532 boolean titleAppended = false;
533 final StringBuilder foreignBuilder = new StringBuilder();
534 final Collection<String> wordForms = new ArrayList<String>();
535 final List<ListSection> listSections = new ArrayList<ListSection>();
539 ListSection lastListSection = null;
541 int currentHeadingDepth = posDepth;
542 while (wikiTokenizer.nextToken() != null) {
543 if (wikiTokenizer.isHeading()) {
544 currentHeadingDepth = wikiTokenizer.headingDepth();
546 if (currentHeadingDepth <= posDepth) {
547 wikiTokenizer.returnToLineStart();
552 if (currentHeadingDepth > posDepth) {
553 // TODO: deal with other neat info sections
557 if (wikiTokenizer.isFunction()) {
558 final String name = wikiTokenizer.functionName();
559 final List<String> args = wikiTokenizer.functionPositionArgs();
560 final Map<String,String> namedArgs = wikiTokenizer.functionNamedArgs();
561 // First line is generally a repeat of the title with some extra information.
562 // We need to build up the left side (foreign text, tokens) separately from the
563 // right side (English). The left-side may get paired with multiple right sides.
564 // The left side should get filed under every form of the word in question (singular, plural).
566 // For verbs, the conjugation comes later on in a deeper section.
567 // Ideally, we'd want to file every English entry with the verb
568 // under every verb form coming from the conjugation.
569 // Ie. under "fa": see: "make :: fare" and "do :: fare"
570 // But then where should we put the conjugation table?
571 // 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!)
572 // for the conjugation table from "fa".
573 // Would like to be able to link to a lang#token.
574 if (isGender(name)) {
575 appendGender(foreignBuilder, name, args);
576 } else if (name.equals("wikipedia")) {
577 namedArgs.remove("lang");
578 if (args.size() > 1 || !namedArgs.isEmpty()) {
580 foreignBuilder.append(wikiTokenizer.token());
581 } else if (args.size() == 1) {
582 foreignBuilder.append(wikiTokenizer.token());
584 //foreignBuilder.append(title);
586 } else if (name.equals("attention") || name.equals("zh-attention")) {
587 // See: http://en.wiktionary.org/wiki/Template:attention
589 } else if (name.equals("infl")) {
590 // See: http://en.wiktionary.org/wiki/Template:infl
591 final String langCode = get(args, 0);
592 String head = namedArgs.remove("head");
594 head = namedArgs.remove("title"); // Bug
599 head = WikiTokenizer.toPlainText(head);
601 titleAppended = true;
603 namedArgs.keySet().removeAll(USELESS_WIKI_ARGS);
605 final String tr = namedArgs.remove("tr");
606 String g = namedArgs.remove("g");
608 g = namedArgs.remove("gender");
610 final String g2 = namedArgs.remove("g2");
611 final String g3 = namedArgs.remove("g3");
613 foreignBuilder.append(head);
616 foreignBuilder.append(" {").append(g);
618 foreignBuilder.append("|").append(g2);
621 foreignBuilder.append("|").append(g3);
623 foreignBuilder.append("}");
627 foreignBuilder.append(String.format(TRANSLITERATION_FORMAT, tr));
631 final String pos = get(args, 1);
633 foreignBuilder.append(" (").append(pos).append(")");
635 for (int i = 2; i < args.size(); i += 2) {
636 final String inflName = get(args, i);
637 final String inflValue = get(args, i + 1);
638 foreignBuilder.append(", ").append(WikiTokenizer.toPlainText(inflName));
639 if (inflValue != null && inflValue.length() > 0) {
640 foreignBuilder.append(": ").append(WikiTokenizer.toPlainText(inflValue));
641 wordForms.add(inflValue);
644 for (final String key : namedArgs.keySet()) {
645 final String value = WikiTokenizer.toPlainText(namedArgs.get(key));
646 foreignBuilder.append(" ").append(key).append("=").append(value);
647 wordForms.add(value);
649 } else if (name.equals("it-noun")) {
650 titleAppended = true;
651 final String base = get(args, 0);
652 final String gender = get(args, 1);
653 final String singular = base + get(args, 2);
654 final String plural = base + get(args, 3);
655 foreignBuilder.append(String.format(" %s {%s}, %s {pl}", singular, gender, plural, plural));
656 wordForms.add(singular);
657 wordForms.add(plural);
658 if (!namedArgs.isEmpty() || args.size() > 4) {
659 LOG.warning("Invalid it-noun: " + wikiTokenizer.token());
661 } else if (name.equals("it-proper noun")) {
662 foreignBuilder.append(wikiTokenizer.token());
663 } else if (name.equals("it-adj")) {
664 foreignBuilder.append(wikiTokenizer.token());
665 } else if (name.startsWith("it-conj")) {
666 if (name.equals("it-conj-are")) {
667 itConjAre(args, namedArgs);
668 } else if (name.equals("it-conj-ere")) {
669 } else if (name.equals("it-conj-ire")) {
671 LOG.warning("Unknown conjugation: " + wikiTokenizer.token());
675 foreignBuilder.append(wikiTokenizer.token());
676 // LOG.warning("Unknown function: " + wikiTokenizer.token());
678 } else if (wikiTokenizer.isListItem()) {
679 final String prefix = wikiTokenizer.listItemPrefix();
680 if (lastListSection != null &&
681 prefix.startsWith(lastListSection.firstPrefix) &&
682 prefix.length() > lastListSection.firstPrefix.length()) {
683 lastListSection.nextPrefixes.add(prefix);
684 lastListSection.nextLines.add(wikiTokenizer.listItemWikiText());
686 lastListSection = new ListSection(prefix, wikiTokenizer.listItemWikiText());
687 listSections.add(lastListSection);
689 } else if (lastListSection != null) {
690 // Don't append anything after the lists, because there's crap.
691 } else if (wikiTokenizer.isWikiLink()) {
693 foreignBuilder.append(wikiTokenizer.wikiLinkText());
695 } else if (wikiTokenizer.isPlainText()) {
697 foreignBuilder.append(wikiTokenizer.token());
699 } else if (wikiTokenizer.isMarkup() || wikiTokenizer.isNewline() || wikiTokenizer.isComment()) {
702 LOG.warning("Unexpected token: " + wikiTokenizer.token());
707 // Here's where we exit.
708 // Should we make an entry even if there are no foreign list items?
709 String foreign = foreignBuilder.toString().trim();
710 if (!titleAppended && !foreign.toLowerCase().startsWith(title.toLowerCase())) {
711 foreign = String.format("%s %s", title, foreign);
713 if (!langPattern.matcher(lang).matches()) {
714 foreign = String.format("(%s) %s", lang, foreign);
716 for (final ListSection listSection : listSections) {
717 doForeignListItem(foreign, title, wordForms, listSection);
723 static final Pattern UNINDEXED_WIKI_TEXT = Pattern.compile(
724 "(first|second|third)-person (singular|plural)|" +
729 // Might only want to remove "lang" if it's equal to "zh", for example.
730 static final Set<String> USELESS_WIKI_ARGS = new LinkedHashSet<String>(Arrays.asList("lang", "sc", "sort", "cat"));
732 private void doForeignListItem(final String foreignText, String title, final Collection<String> forms, final ListSection listSection) {
734 final String prefix = listSection.firstPrefix;
735 if (prefix.length() > 1) {
736 // Could just get looser and say that any prefix longer than first is a sublist.
737 LOG.warning("Prefix too long: " + listSection);
741 final PairEntry pairEntry = new PairEntry();
742 final IndexedEntry indexedEntry = new IndexedEntry(pairEntry);
744 final StringBuilder englishBuilder = new StringBuilder();
746 final String mainLine = listSection.firstLine;
747 final WikiTokenizer englishTokenizer = new WikiTokenizer(mainLine, false);
748 while (englishTokenizer.nextToken() != null) {
749 // TODO handle form of....
750 if (englishTokenizer.isPlainText()) {
751 englishBuilder.append(englishTokenizer.token());
752 enIndexBuilder.addEntryWithString(indexedEntry, englishTokenizer.token(), EntryTypeName.WIKTIONARY_ENGLISH_DEF);
753 } else if (englishTokenizer.isWikiLink()) {
754 final String text = englishTokenizer.wikiLinkText();
755 final String link = englishTokenizer.wikiLinkDest();
757 if (link.contains("#English")) {
758 englishBuilder.append(text);
759 enIndexBuilder.addEntryWithString(indexedEntry, text, EntryTypeName.WIKTIONARY_ENGLISH_DEF_WIKI_LINK);
760 } else if (link.contains("#") && this.langPattern.matcher(link).find()) {
761 englishBuilder.append(text);
762 foreignIndexBuilder.addEntryWithString(indexedEntry, text, EntryTypeName.WIKTIONARY_ENGLISH_DEF_OTHER_LANG);
763 } else if (link.equals("plural")) {
764 englishBuilder.append(text);
766 //LOG.warning("Special link: " + englishTokenizer.token());
767 enIndexBuilder.addEntryWithString(indexedEntry, text, EntryTypeName.WIKTIONARY_ENGLISH_DEF_WIKI_LINK);
768 englishBuilder.append(text);
772 englishBuilder.append(text);
773 if (!UNINDEXED_WIKI_TEXT.matcher(text).find()) {
774 enIndexBuilder.addEntryWithString(indexedEntry, text, EntryTypeName.WIKTIONARY_ENGLISH_DEF_WIKI_LINK);
777 } else if (englishTokenizer.isFunction()) {
778 final String name = englishTokenizer.functionName();
779 final List<String> args = englishTokenizer.functionPositionArgs();
780 final Map<String,String> namedArgs = englishTokenizer.functionNamedArgs();
783 name.equals("form of") ||
784 name.contains("conjugation of") ||
785 name.contains("participle of") ||
786 name.contains("gerund of") ||
787 name.contains("feminine of") ||
788 name.contains("plural of")) {
789 String formName = name;
790 if (name.equals("form of")) {
791 formName = remove(args, 0, null);
793 if (formName == null) {
794 LOG.warning("Missing form name: " + title);
795 formName = "form of";
797 String baseForm = get(args, 1, "");
798 if ("".equals(baseForm)) {
799 baseForm = get(args, 0, null);
802 remove(args, 0, null);
804 namedArgs.keySet().removeAll(USELESS_WIKI_ARGS);
805 WikiTokenizer.appendFunction(englishBuilder.append("{"), formName, args, namedArgs).append("}");
806 if (baseForm != null) {
807 foreignIndexBuilder.addEntryWithString(indexedEntry, baseForm, EntryTypeName.WIKTIONARY_BASE_FORM_SINGLE, EntryTypeName.WIKTIONARY_BASE_FORM_MULTI);
809 // null baseForm happens in Danish.
810 LOG.warning("Null baseform: " + title);
813 namedArgs.keySet().removeAll(USELESS_WIKI_ARGS);
814 if (args.size() == 0 && namedArgs.isEmpty()) {
815 englishBuilder.append("{").append(name).append("}");
817 WikiTokenizer.appendFunction(englishBuilder.append("{{"), name, args, namedArgs).append("}}");
819 // LOG.warning("Unexpected function: " + englishTokenizer.token());
822 if (englishTokenizer.isComment() || englishTokenizer.isMarkup()) {
824 LOG.warning("Unexpected definition type: " + englishTokenizer.token());
829 final String english = trim(englishBuilder.toString());
830 if (english.length() > 0) {
831 final Pair pair = new Pair(english, trim(foreignText), this.swap);
832 pairEntry.pairs.add(pair);
833 foreignIndexBuilder.addEntryWithString(indexedEntry, title, EntryTypeName.WIKTIONARY_TITLE_SINGLE, EntryTypeName.WIKTIONARY_TITLE_MULTI);
834 for (final String form : forms) {
835 foreignIndexBuilder.addEntryWithString(indexedEntry, form, EntryTypeName.WIKTIONARY_INFLECTD_FORM_SINGLE, EntryTypeName.WIKTIONARY_INFLECTED_FORM_MULTI);
840 String lastForeign = null;
841 for (int i = 0; i < listSection.nextPrefixes.size(); ++i) {
842 final String nextPrefix = listSection.nextPrefixes.get(i);
843 final String nextLine = listSection.nextLines.get(i);
844 int dash = nextLine.indexOf("—");
847 dash = nextLine.indexOf("—");
851 dash = nextLine.indexOf(" - ");
855 if ((nextPrefix.equals("#:") || nextPrefix.equals("##:")) && dash != -1) {
856 final String foreignEx = nextLine.substring(0, dash);
857 final String englishEx = nextLine.substring(dash + mdashLen);
858 final Pair pair = new Pair(formatAndIndexExampleString(englishEx, enIndexBuilder, indexedEntry), formatAndIndexExampleString(foreignEx, foreignIndexBuilder, indexedEntry), swap);
859 if (pair.lang1 != "--" && pair.lang1 != "--") {
860 pairEntry.pairs.add(pair);
863 } else if (nextPrefix.equals("#:") || nextPrefix.equals("##:")){
864 final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
865 lastForeign = nextLine;
866 if (pair.lang1 != "--" && pair.lang1 != "--") {
867 pairEntry.pairs.add(pair);
869 } else if (nextPrefix.equals("#::") || nextPrefix.equals("#**")) {
870 if (lastForeign != null && pairEntry.pairs.size() > 0) {
871 pairEntry.pairs.remove(pairEntry.pairs.size() - 1);
872 final Pair pair = new Pair(formatAndIndexExampleString(nextLine, enIndexBuilder, indexedEntry), formatAndIndexExampleString(lastForeign, foreignIndexBuilder, indexedEntry), swap);
873 if (pair.lang1 != "--" || pair.lang2 != "--") {
874 pairEntry.pairs.add(pair);
878 LOG.warning("TODO: English example with no foreign: " + title + ", " + nextLine);
879 final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
880 if (pair.lang1 != "--" || pair.lang2 != "--") {
881 pairEntry.pairs.add(pair);
884 } else if (nextPrefix.equals("#*")) {
885 // Can't really index these.
886 final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
887 lastForeign = nextLine;
888 if (pair.lang1 != "--" || pair.lang2 != "--") {
889 pairEntry.pairs.add(pair);
891 } else if (nextPrefix.equals("#::*") || nextPrefix.equals("##") || nextPrefix.equals("#*:") || nextPrefix.equals("#:*") || true) {
892 final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
893 if (pair.lang1 != "--" || pair.lang2 != "--") {
894 pairEntry.pairs.add(pair);
902 private String formatAndIndexExampleString(final String example, final IndexBuilder indexBuilder, final IndexedEntry indexedEntry) {
903 final WikiTokenizer wikiTokenizer = new WikiTokenizer(example, false);
904 final StringBuilder builder = new StringBuilder();
905 boolean insideTripleQuotes = false;
906 while (wikiTokenizer.nextToken() != null) {
907 if (wikiTokenizer.isPlainText()) {
908 builder.append(wikiTokenizer.token());
909 if (indexBuilder != null) {
910 indexBuilder.addEntryWithString(indexedEntry, wikiTokenizer.token(), EntryTypeName.WIKTIONARY_EXAMPLE);
912 } else if (wikiTokenizer.isWikiLink()) {
913 final String text = wikiTokenizer.wikiLinkText().replaceAll("'", "");
914 builder.append(text);
915 if (indexBuilder != null) {
916 indexBuilder.addEntryWithString(indexedEntry, text, EntryTypeName.WIKTIONARY_EXAMPLE);
918 } else if (wikiTokenizer.isFunction()) {
919 builder.append(wikiTokenizer.token());
920 } else if (wikiTokenizer.isMarkup()) {
921 if (wikiTokenizer.token().equals("'''")) {
922 insideTripleQuotes = !insideTripleQuotes;
924 } else if (wikiTokenizer.isComment() || wikiTokenizer.isNewline()) {
927 LOG.warning("unexpected token: " + wikiTokenizer.token());
930 final String result = trim(builder.toString());
931 return result.length() > 0 ? result : "--";
935 private void itConjAre(List<String> args, Map<String, String> namedArgs) {
936 final String base = args.get(0);
937 final String aux = args.get(1);
939 putIfMissing(namedArgs, "inf", base + "are");
940 putIfMissing(namedArgs, "aux", aux);
941 putIfMissing(namedArgs, "ger", base + "ando");
942 putIfMissing(namedArgs, "presp", base + "ante");
943 putIfMissing(namedArgs, "pastp", base + "ato");
945 putIfMissing(namedArgs, "pres1s", base + "o");
946 putIfMissing(namedArgs, "pres2s", base + "i");
947 putIfMissing(namedArgs, "pres3s", base + "a");
948 putIfMissing(namedArgs, "pres1p", base + "iamo");
949 putIfMissing(namedArgs, "pres2p", base + "ate");
950 putIfMissing(namedArgs, "pres3p", base + "ano");
952 putIfMissing(namedArgs, "imperf1s", base + "avo");
953 putIfMissing(namedArgs, "imperf2s", base + "avi");
954 putIfMissing(namedArgs, "imperf3s", base + "ava");
955 putIfMissing(namedArgs, "imperf1p", base + "avamo");
956 putIfMissing(namedArgs, "imperf2p", base + "avate");
957 putIfMissing(namedArgs, "imperf3p", base + "avano");
959 putIfMissing(namedArgs, "prem1s", base + "ai");
960 putIfMissing(namedArgs, "prem2s", base + "asti");
961 putIfMissing(namedArgs, "prem3s", base + "ò");
962 putIfMissing(namedArgs, "prem1p", base + "ammo");
963 putIfMissing(namedArgs, "prem2p", base + "aste");
964 putIfMissing(namedArgs, "prem3p", base + "arono");
966 putIfMissing(namedArgs, "fut1s", base + "erò");
967 putIfMissing(namedArgs, "fut2s", base + "erai");
968 putIfMissing(namedArgs, "fut3s", base + "erà");
969 putIfMissing(namedArgs, "fut1p", base + "eremo");
970 putIfMissing(namedArgs, "fut2p", base + "erete");
971 putIfMissing(namedArgs, "fut3p", base + "eranno");
973 putIfMissing(namedArgs, "cond1s", base + "erei");
974 putIfMissing(namedArgs, "cond2s", base + "eresti");
975 putIfMissing(namedArgs, "cond3s", base + "erebbe");
976 putIfMissing(namedArgs, "cond1p", base + "eremmo");
977 putIfMissing(namedArgs, "cond2p", base + "ereste");
978 putIfMissing(namedArgs, "cond3p", base + "erebbero");
979 // Subjunctive / congiuntivo
980 putIfMissing(namedArgs, "sub123s", base + "i");
981 putIfMissing(namedArgs, "sub1p", base + "iamo");
982 putIfMissing(namedArgs, "sub2p", base + "iate");
983 putIfMissing(namedArgs, "sub3p", base + "ino");
984 // Imperfect subjunctive
985 putIfMissing(namedArgs, "impsub12s", base + "assi");
986 putIfMissing(namedArgs, "impsub3s", base + "asse");
987 putIfMissing(namedArgs, "impsub1p", base + "assimo");
988 putIfMissing(namedArgs, "impsub2p", base + "aste");
989 putIfMissing(namedArgs, "impsub3p", base + "assero");
991 putIfMissing(namedArgs, "imp2s", base + "a");
992 putIfMissing(namedArgs, "imp3s", base + "i");
993 putIfMissing(namedArgs, "imp1p", base + "iamo");
994 putIfMissing(namedArgs, "imp2p", base + "ate");
995 putIfMissing(namedArgs, "imp3p", base + "ino");
998 itConj(args, namedArgs);
1002 private void itConj(List<String> args, Map<String, String> namedArgs) {
1003 // TODO Auto-generated method stub
1008 private static void putIfMissing(final Map<String, String> namedArgs, final String key,
1009 final String value) {
1010 final String oldValue = namedArgs.get(key);
1011 if (oldValue == null || oldValue.length() == 0) {
1012 namedArgs.put(key, value);
1016 // TODO: check how ='' and =| are manifested....
1017 // TODO: get this right in -are
1018 private static void putOrNullify(final Map<String, String> namedArgs, final String key,
1019 final String value) {
1020 final String oldValue = namedArgs.get(key);
1021 if (oldValue == null/* || oldValue.length() == 0*/) {
1022 namedArgs.put(key, value);
1024 if (oldValue.equals("''")) {
1025 namedArgs.put(key, "");
1030 static final Pattern whitespace = Pattern.compile("\\s+");
1031 static String trim(final String s) {
1032 return whitespace.matcher(s).replaceAll(" ").trim();