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;
39 import com.hughes.util.ListUtil;
41 public class EnWiktionaryXmlParser {
43 private static final String TRANSLITERATION_FORMAT = " (tr. %s)";
45 static final Logger LOG = Logger.getLogger(EnWiktionaryXmlParser.class.getName());
47 // TODO: process {{ttbc}} lines
49 static final Pattern partOfSpeechHeader = Pattern.compile(
50 "Noun|Verb|Adjective|Adverb|Pronoun|Conjunction|Interjection|" +
51 "Preposition|Proper noun|Article|Prepositional phrase|Acronym|" +
52 "Abbreviation|Initialism|Contraction|Prefix|Suffix|Symbol|Letter|" +
53 "Ligature|Idiom|Phrase|\\{\\{acronym\\}\\}|\\{\\{initialism\\}\\}|" +
54 // These are @deprecated:
55 "Noun form|Verb form|Adjective form|Nominal phrase|Noun phrase|" +
56 "Verb phrase|Transitive verb|Intransitive verb|Reflexive verb|" +
57 // These are extras I found:
58 "Determiner|Numeral|Number|Cardinal number|Ordinal number|Proverb|" +
59 "Particle|Interjection|Pronominal adverb" +
60 "Han character|Hanzi|Hanja|Kanji|Katakana character|Syllable");
62 final IndexBuilder enIndexBuilder;
63 final IndexBuilder foreignIndexBuilder;
64 final Pattern langPattern;
65 final Pattern langCodePattern;
68 // State used while parsing.
71 public EnWiktionaryXmlParser(final IndexBuilder enIndexBuilder, final IndexBuilder otherIndexBuilder, final Pattern langPattern, final Pattern langCodePattern, final boolean swap) {
72 this.enIndexBuilder = enIndexBuilder;
73 this.foreignIndexBuilder = otherIndexBuilder;
74 this.langPattern = langPattern;
75 this.langCodePattern = langCodePattern;
80 public void parse(final File file, final int pageLimit) throws IOException {
82 final DataInputStream dis = new DataInputStream(new BufferedInputStream(new FileInputStream(file)));
84 if (pageLimit >= 0 && pageCount >= pageLimit) {
89 title = dis.readUTF();
90 } catch (EOFException e) {
91 LOG.warning("Error reading split!");
95 final String heading = dis.readUTF();
96 final int bytesLength = dis.readInt();
97 final byte[] bytes = new byte[bytesLength];
99 final String text = new String(bytes, "UTF8");
101 parseSection(heading, text);
104 if (pageCount % 1000 == 0) {
105 LOG.info("pageCount=" + pageCount);
110 private void parseSection(String heading, final String text) {
111 if (title.startsWith("Wiktionary:") ||
112 title.startsWith("Template:") ||
113 title.startsWith("Appendix:") ||
114 title.startsWith("Category:") ||
115 title.startsWith("Index:") ||
116 title.startsWith("MediaWiki:") ||
117 title.startsWith("TransWiki:") ||
118 title.startsWith("Citations:") ||
119 title.startsWith("Concordance:") ||
120 title.startsWith("Help:")) {
124 heading = heading.replaceAll("=", "").trim();
125 if (heading.equals("English")) {
127 } else if (langPattern.matcher(heading).find()){
128 doForeignWord(heading, text);
133 // -------------------------------------------------------------------------
135 private void doEnglishWord(String text) {
140 final WikiTokenizer wikiTokenizer = new WikiTokenizer(text);
141 while (wikiTokenizer.nextToken() != null) {
143 if (wikiTokenizer.isHeading()) {
144 final String headerName = wikiTokenizer.headingWikiText();
146 if (wikiTokenizer.headingDepth() <= posDepth) {
151 if (partOfSpeechHeader.matcher(headerName).matches()) {
152 posDepth = wikiTokenizer.headingDepth();
153 pos = wikiTokenizer.headingWikiText();
154 // TODO: if we're inside the POS section, we should handle the first title line...
156 } else if (headerName.equals("Translations")) {
158 LOG.warning("Translations without POS: " + title);
160 doTranslations(wikiTokenizer, pos);
161 } else if (headerName.equals("Pronunciation")) {
162 //doPronunciation(wikiLineReader);
164 } else if (wikiTokenizer.isFunction()) {
165 final String name = wikiTokenizer.functionName();
166 if (name.equals("head")) {
167 LOG.warning("{{head}} without POS: " + title);
173 final AppendAndIndexWikiCallback appendAndIndexWikiCallback = new AppendAndIndexWikiCallback(this);
175 private void doTranslations(final WikiTokenizer wikiTokenizer, final String pos) {
176 if (title.equals("absolutely")) {
177 //System.out.println();
180 String topLevelLang = null;
182 boolean done = false;
183 while (wikiTokenizer.nextToken() != null) {
184 if (wikiTokenizer.isHeading()) {
185 wikiTokenizer.returnToLineStart();
192 // Check whether we care about this line:
194 if (wikiTokenizer.isFunction()) {
195 final String functionName = wikiTokenizer.functionName();
196 final List<String> positionArgs = wikiTokenizer.functionPositionArgs();
198 if (functionName.equals("trans-top")) {
200 if (wikiTokenizer.functionPositionArgs().size() >= 1) {
201 sense = positionArgs.get(0);
202 // TODO: could emphasize words in [[brackets]] inside sense.
203 sense = WikiTokenizer.toPlainText(sense);
204 //LOG.info("Sense: " + sense);
206 } else if (functionName.equals("trans-bottom")) {
208 } else if (functionName.equals("trans-mid")) {
209 } else if (functionName.equals("trans-see")) {
210 // TODO: would also be nice...
211 } else if (functionName.startsWith("picdic")) {
212 } else if (functionName.startsWith("checktrans")) {
214 } else if (functionName.startsWith("ttbc")) {
215 wikiTokenizer.nextLine();
216 // TODO: would be great to handle ttbc
217 // TODO: Check this: done = true;
219 LOG.warning("Unexpected translation wikifunction: " + wikiTokenizer.token() + ", title=" + title);
221 } else if (wikiTokenizer.isListItem()) {
222 final String line = wikiTokenizer.listItemWikiText();
223 // This line could produce an output...
225 if (line.contains("ich hoan dich gear")) {
226 //System.out.println();
229 // First strip the language and check whether it matches.
230 // And hold onto it for sub-lines.
231 final int colonIndex = line.indexOf(":");
232 if (colonIndex == -1) {
236 final String lang = trim(WikiTokenizer.toPlainText(line.substring(0, colonIndex)));
237 final boolean appendLang;
238 if (wikiTokenizer.listItemPrefix().length() == 1) {
240 final boolean thisFind = langPattern.matcher(lang).find();
244 appendLang = !langPattern.matcher(lang).matches();
245 } else if (topLevelLang == null) {
248 // Two-level -- the only way we won't append is if this second level matches exactly.
249 if (!langPattern.matcher(lang).matches() && !langPattern.matcher(topLevelLang).find()) {
252 appendLang = !langPattern.matcher(lang).matches();
255 String rest = line.substring(colonIndex + 1).trim();
256 if (rest.length() > 0) {
257 doTranslationLine(line, appendLang ? lang : null, pos, sense, rest);
260 } else if (wikiTokenizer.remainderStartsWith("''See''")) {
261 wikiTokenizer.nextLine();
262 LOG.fine("Skipping See line: " + wikiTokenizer.token());
263 } else if (wikiTokenizer.isWikiLink()) {
264 final String wikiLink = wikiTokenizer.wikiLinkText();
265 if (wikiLink.contains(":") && wikiLink.contains(title)) {
266 } else if (wikiLink.contains("Category:")) {
268 LOG.warning("Unexpected wikiLink: " + wikiTokenizer.token() + ", title=" + title);
270 } else if (wikiTokenizer.isNewline() || wikiTokenizer.isMarkup() || wikiTokenizer.isComment()) {
272 final String token = wikiTokenizer.token();
273 if (token.equals("----")) {
275 LOG.warning("Unexpected translation token: " + wikiTokenizer.token() + ", title=" + title);
282 private void doTranslationLine(final String line, final String lang, final String pos, final String sense, final String rest) {
283 // Good chance we'll actually file this one...
284 final PairEntry pairEntry = new PairEntry();
285 final IndexedEntry indexedEntry = new IndexedEntry(pairEntry);
287 final StringBuilder foreignText = new StringBuilder();
288 appendAndIndexWikiCallback.reset(foreignText, indexedEntry);
289 appendAndIndexWikiCallback.functionCallbacks.clear();
290 appendAndIndexWikiCallback.functionCallbacks.putAll(FunctionCallbacksDefault.DEFAULT);
291 appendAndIndexWikiCallback.dispatch(rest, foreignIndexBuilder, EntryTypeName.WIKTIONARY_TRANSLATION_OTHER_TEXT);
293 if (foreignText.length() == 0) {
294 LOG.warning("Empty foreignText: " + line);
299 foreignText.insert(0, String.format("(%s) ", lang));
302 StringBuilder englishText = new StringBuilder();
304 englishText.append(title);
306 englishText.append(" (").append(sense).append(")");
307 enIndexBuilder.addEntryWithString(indexedEntry, sense, EntryTypeName.WIKTIONARY_TRANSLATION_SENSE);
310 englishText.append(" (").append(pos.toLowerCase()).append(")");
312 enIndexBuilder.addEntryWithString(indexedEntry, title, EntryTypeName.WIKTIONARY_TITLE_MULTI);
314 final Pair pair = new Pair(trim(englishText.toString()), trim(foreignText.toString()), swap);
315 pairEntry.pairs.add(pair);
316 if (!pairsAdded.add(pair.toString())) {
317 LOG.warning("Duplicate pair: " + pair.toString());
319 if (pair.toString().equals("libero {m} :: free (adjective)")) {
320 System.out.println();
326 Set<String> pairsAdded = new LinkedHashSet<String>();
328 // -------------------------------------------------------------------------
330 private void doForeignWord(final String lang, final String text) {
331 final WikiTokenizer wikiTokenizer = new WikiTokenizer(text);
332 while (wikiTokenizer.nextToken() != null) {
333 if (wikiTokenizer.isHeading()) {
334 final String headingName = wikiTokenizer.headingWikiText();
335 if (headingName.equals("Translations")) {
336 LOG.warning("Translations not in English section: " + title);
337 } else if (headingName.equals("Pronunciation")) {
338 //doPronunciation(wikiLineReader);
339 } else if (partOfSpeechHeader.matcher(headingName).matches()) {
340 doForeignPartOfSpeech(lang, headingName, wikiTokenizer.headingDepth(), wikiTokenizer);
347 static final class ListSection {
348 final String firstPrefix;
349 final String firstLine;
350 final List<String> nextPrefixes = new ArrayList<String>();
351 final List<String> nextLines = new ArrayList<String>();
353 public ListSection(String firstPrefix, String firstLine) {
354 this.firstPrefix = firstPrefix;
355 this.firstLine = firstLine;
359 public String toString() {
360 return firstPrefix + firstLine + "{ " + nextPrefixes + "}";
365 int foreignCount = 0;
366 private void doForeignPartOfSpeech(final String lang, String posHeading, final int posDepth, WikiTokenizer wikiTokenizer) {
367 if (++foreignCount % 1000 == 0) {
368 LOG.info("***" + lang + ", " + title + ", pos=" + posHeading + ", foreignCount=" + foreignCount);
370 if (title.equals("moro")) {
371 System.out.println();
374 boolean titleAppended = false;
375 final StringBuilder foreignBuilder = new StringBuilder();
376 final Collection<String> wordForms = new ArrayList<String>();
377 final List<ListSection> listSections = new ArrayList<ListSection>();
381 ListSection lastListSection = null;
383 int currentHeadingDepth = posDepth;
384 while (wikiTokenizer.nextToken() != null) {
385 if (wikiTokenizer.isHeading()) {
386 currentHeadingDepth = wikiTokenizer.headingDepth();
388 if (currentHeadingDepth <= posDepth) {
389 wikiTokenizer.returnToLineStart();
394 if (currentHeadingDepth > posDepth) {
395 // TODO: deal with other neat info sections
399 if (wikiTokenizer.isFunction()) {
400 final String name = wikiTokenizer.functionName();
401 final List<String> args = wikiTokenizer.functionPositionArgs();
402 final Map<String,String> namedArgs = wikiTokenizer.functionNamedArgs();
403 // First line is generally a repeat of the title with some extra information.
404 // We need to build up the left side (foreign text, tokens) separately from the
405 // right side (English). The left-side may get paired with multiple right sides.
406 // The left side should get filed under every form of the word in question (singular, plural).
408 // For verbs, the conjugation comes later on in a deeper section.
409 // Ideally, we'd want to file every English entry with the verb
410 // under every verb form coming from the conjugation.
411 // Ie. under "fa": see: "make :: fare" and "do :: fare"
412 // But then where should we put the conjugation table?
413 // 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!)
414 // for the conjugation table from "fa".
415 // Would like to be able to link to a lang#token.
416 if (FunctionCallbacksDefault.DEFAULT.get(name) instanceof FunctionCallbacksDefault.Gender) {
418 appendAndIndexWikiCallback.reset(foreignBuilder, null);
419 FunctionCallbacksDefault.DEFAULT.get(name).onWikiFunction(wikiTokenizer, name, args, namedArgs, this, appendAndIndexWikiCallback);
420 } else if (name.equals("wikipedia")) {
421 namedArgs.remove("lang");
422 if (args.size() > 1 || !namedArgs.isEmpty()) {
424 foreignBuilder.append(wikiTokenizer.token());
425 } else if (args.size() == 1) {
426 foreignBuilder.append(wikiTokenizer.token());
428 //foreignBuilder.append(title);
430 } else if (name.equals("attention") || name.equals("zh-attention")) {
431 // See: http://en.wiktionary.org/wiki/Template:attention
433 } else if (name.equals("infl") || name.equals("head")) {
434 // See: http://en.wiktionary.org/wiki/Template:infl
435 final String langCode = ListUtil.get(args, 0);
436 String head = namedArgs.remove("head");
438 head = namedArgs.remove("title"); // Bug
443 head = WikiTokenizer.toPlainText(head);
445 titleAppended = true;
447 namedArgs.keySet().removeAll(USELESS_WIKI_ARGS);
449 final String tr = namedArgs.remove("tr");
450 String g = namedArgs.remove("g");
452 g = namedArgs.remove("gender");
454 final String g2 = namedArgs.remove("g2");
455 final String g3 = namedArgs.remove("g3");
457 foreignBuilder.append(head);
460 foreignBuilder.append(" {").append(g);
462 foreignBuilder.append("|").append(g2);
465 foreignBuilder.append("|").append(g3);
467 foreignBuilder.append("}");
471 foreignBuilder.append(String.format(TRANSLITERATION_FORMAT, tr));
475 final String pos = ListUtil.get(args, 1);
477 foreignBuilder.append(" (").append(pos).append(")");
479 for (int i = 2; i < args.size(); i += 2) {
480 final String inflName = ListUtil.get(args, i);
481 final String inflValue = ListUtil.get(args, i + 1);
482 foreignBuilder.append(", ").append(WikiTokenizer.toPlainText(inflName));
483 if (inflValue != null && inflValue.length() > 0) {
484 foreignBuilder.append(": ").append(WikiTokenizer.toPlainText(inflValue));
485 wordForms.add(inflValue);
488 for (final String key : namedArgs.keySet()) {
489 final String value = WikiTokenizer.toPlainText(namedArgs.get(key));
490 foreignBuilder.append(" ").append(key).append("=").append(value);
491 wordForms.add(value);
493 } else if (name.equals("it-noun")) {
494 titleAppended = true;
495 final String base = ListUtil.get(args, 0);
496 final String gender = ListUtil.get(args, 1);
497 final String singular = base + ListUtil.get(args, 2, null);
498 final String plural = base + ListUtil.get(args, 3, null);
499 foreignBuilder.append(String.format(" %s {%s}, %s {pl}", singular, gender, plural, plural));
500 wordForms.add(singular);
501 wordForms.add(plural);
502 if (!namedArgs.isEmpty() || args.size() > 4) {
503 LOG.warning("Invalid it-noun: " + wikiTokenizer.token());
505 } else if (name.equals("it-proper noun")) {
506 foreignBuilder.append(wikiTokenizer.token());
507 } else if (name.equals("it-adj")) {
508 foreignBuilder.append(wikiTokenizer.token());
509 } else if (name.startsWith("it-conj")) {
510 if (name.equals("it-conj-are")) {
511 itConjAre(args, namedArgs);
512 } else if (name.equals("it-conj-ere")) {
513 } else if (name.equals("it-conj-ire")) {
515 LOG.warning("Unknown conjugation: " + wikiTokenizer.token());
519 foreignBuilder.append(wikiTokenizer.token());
520 // LOG.warning("Unknown function: " + wikiTokenizer.token());
522 } else if (wikiTokenizer.isListItem()) {
523 final String prefix = wikiTokenizer.listItemPrefix();
524 if (lastListSection != null &&
525 prefix.startsWith(lastListSection.firstPrefix) &&
526 prefix.length() > lastListSection.firstPrefix.length()) {
527 lastListSection.nextPrefixes.add(prefix);
528 lastListSection.nextLines.add(wikiTokenizer.listItemWikiText());
530 lastListSection = new ListSection(prefix, wikiTokenizer.listItemWikiText());
531 listSections.add(lastListSection);
533 } else if (lastListSection != null) {
534 // Don't append anything after the lists, because there's crap.
535 } else if (wikiTokenizer.isWikiLink()) {
537 foreignBuilder.append(wikiTokenizer.wikiLinkText());
539 } else if (wikiTokenizer.isPlainText()) {
541 foreignBuilder.append(wikiTokenizer.token());
543 } else if (wikiTokenizer.isMarkup() || wikiTokenizer.isNewline() || wikiTokenizer.isComment()) {
546 LOG.warning("Unexpected token: " + wikiTokenizer.token());
551 // Here's where we exit.
552 // Should we make an entry even if there are no foreign list items?
553 String foreign = foreignBuilder.toString().trim();
554 if (!titleAppended && !foreign.toLowerCase().startsWith(title.toLowerCase())) {
555 foreign = String.format("%s %s", title, foreign);
557 if (!langPattern.matcher(lang).matches()) {
558 foreign = String.format("(%s) %s", lang, foreign);
560 for (final ListSection listSection : listSections) {
561 doForeignListItem(foreign, title, wordForms, listSection);
567 static final Pattern UNINDEXED_WIKI_TEXT = Pattern.compile(
568 "(first|second|third)-person (singular|plural)|" +
573 // Might only want to remove "lang" if it's equal to "zh", for example.
574 static final Set<String> USELESS_WIKI_ARGS = new LinkedHashSet<String>(
582 private void doForeignListItem(final String foreignText, String title, final Collection<String> forms, final ListSection listSection) {
584 final String prefix = listSection.firstPrefix;
585 if (prefix.length() > 1) {
586 // Could just get looser and say that any prefix longer than first is a sublist.
587 LOG.warning("Prefix too long: " + listSection);
591 final PairEntry pairEntry = new PairEntry();
592 final IndexedEntry indexedEntry = new IndexedEntry(pairEntry);
594 final StringBuilder englishBuilder = new StringBuilder();
596 final String mainLine = listSection.firstLine;
597 final WikiTokenizer englishTokenizer = new WikiTokenizer(mainLine, false);
598 while (englishTokenizer.nextToken() != null) {
599 // TODO handle form of....
600 if (englishTokenizer.isPlainText()) {
601 englishBuilder.append(englishTokenizer.token());
602 enIndexBuilder.addEntryWithStringNoSingle(indexedEntry, englishTokenizer.token(), EntryTypeName.WIKTIONARY_ENGLISH_DEF);
603 } else if (englishTokenizer.isWikiLink()) {
604 final String text = englishTokenizer.wikiLinkText();
605 final String link = englishTokenizer.wikiLinkDest();
607 if (link.contains("#English")) {
608 englishBuilder.append(text);
609 enIndexBuilder.addEntryWithStringNoSingle(indexedEntry, text, EntryTypeName.WIKTIONARY_ENGLISH_DEF_WIKI_LINK);
610 } else if (link.contains("#") && this.langPattern.matcher(link).find()) {
611 englishBuilder.append(text);
612 foreignIndexBuilder.addEntryWithStringNoSingle(indexedEntry, text, EntryTypeName.WIKTIONARY_ENGLISH_DEF_OTHER_LANG);
613 } else if (link.equals("plural")) {
614 englishBuilder.append(text);
616 //LOG.warning("Special link: " + englishTokenizer.token());
617 enIndexBuilder.addEntryWithStringNoSingle(indexedEntry, text, EntryTypeName.WIKTIONARY_ENGLISH_DEF_WIKI_LINK);
618 englishBuilder.append(text);
622 englishBuilder.append(text);
623 if (!UNINDEXED_WIKI_TEXT.matcher(text).find()) {
624 enIndexBuilder.addEntryWithStringNoSingle(indexedEntry, text, EntryTypeName.WIKTIONARY_ENGLISH_DEF_WIKI_LINK);
627 } else if (englishTokenizer.isFunction()) {
628 final String name = englishTokenizer.functionName();
629 final List<String> args = englishTokenizer.functionPositionArgs();
630 final Map<String,String> namedArgs = englishTokenizer.functionNamedArgs();
633 name.equals("form of") ||
634 name.contains("conjugation of") ||
635 name.contains("participle of") ||
636 name.contains("gerund of") ||
637 name.contains("feminine of") ||
638 name.contains("plural of")) {
639 String formName = name;
640 if (name.equals("form of")) {
641 formName = ListUtil.remove(args, 0, null);
643 if (formName == null) {
644 LOG.warning("Missing form name: " + title);
645 formName = "form of";
647 String baseForm = ListUtil.get(args, 1, "");
648 if ("".equals(baseForm)) {
649 baseForm = ListUtil.get(args, 0, null);
650 ListUtil.remove(args, 1, "");
652 ListUtil.remove(args, 0, null);
654 namedArgs.keySet().removeAll(USELESS_WIKI_ARGS);
655 WikiTokenizer.appendFunction(englishBuilder.append("{"), formName, args, namedArgs).append("}");
656 if (baseForm != null) {
657 foreignIndexBuilder.addEntryWithString(indexedEntry, baseForm, EntryTypeName.WIKTIONARY_BASE_FORM_MULTI);
659 // null baseForm happens in Danish.
660 LOG.warning("Null baseform: " + title);
662 } else if (name.equals("l")) {
663 // encodes text in various langs.
665 englishBuilder.append("").append(args.get(1));
666 final String langCode = args.get(0);
667 if ("en".equals(langCode)) {
668 enIndexBuilder.addEntryWithStringNoSingle(indexedEntry, args.get(1), EntryTypeName.WIKTIONARY_ENGLISH_DEF_WIKI_LINK);
670 foreignIndexBuilder.addEntryWithStringNoSingle(indexedEntry, args.get(1), EntryTypeName.WIKTIONARY_ENGLISH_DEF_OTHER_LANG);
672 // TODO: transliteration
674 } else if (name.equals("defn") || name.equals("rfdef")) {
676 // http://en.wiktionary.org/wiki/Wiktionary:Requests_for_deletion/Others#Template:defn
677 // Redundant, used for the same purpose as {{rfdef}}, but this
678 // doesn't produce the "This word needs a definition" text.
679 // Delete or redirect.
681 namedArgs.keySet().removeAll(USELESS_WIKI_ARGS);
682 if (args.size() == 0 && namedArgs.isEmpty()) {
683 englishBuilder.append("{").append(name).append("}");
685 WikiTokenizer.appendFunction(englishBuilder.append("{{"), name, args, namedArgs).append("}}");
687 // LOG.warning("Unexpected function: " + englishTokenizer.token());
690 if (englishTokenizer.isComment() || englishTokenizer.isMarkup()) {
692 LOG.warning("Unexpected definition type: " + englishTokenizer.token());
697 final String english = trim(englishBuilder.toString());
698 if (english.length() > 0) {
699 final Pair pair = new Pair(english, trim(foreignText), this.swap);
700 pairEntry.pairs.add(pair);
701 foreignIndexBuilder.addEntryWithString(indexedEntry, title, EntryTypeName.WIKTIONARY_TITLE_MULTI);
702 for (final String form : forms) {
703 foreignIndexBuilder.addEntryWithString(indexedEntry, form, EntryTypeName.WIKTIONARY_INFLECTED_FORM_MULTI);
708 String lastForeign = null;
709 for (int i = 0; i < listSection.nextPrefixes.size(); ++i) {
710 final String nextPrefix = listSection.nextPrefixes.get(i);
711 final String nextLine = listSection.nextLines.get(i);
712 int dash = nextLine.indexOf("—");
715 dash = nextLine.indexOf("—");
719 dash = nextLine.indexOf(" - ");
723 if ((nextPrefix.equals("#:") || nextPrefix.equals("##:")) && dash != -1) {
724 final String foreignEx = nextLine.substring(0, dash);
725 final String englishEx = nextLine.substring(dash + mdashLen);
726 final Pair pair = new Pair(formatAndIndexExampleString(englishEx, enIndexBuilder, indexedEntry), formatAndIndexExampleString(foreignEx, foreignIndexBuilder, indexedEntry), swap);
727 if (pair.lang1 != "--" && pair.lang1 != "--") {
728 pairEntry.pairs.add(pair);
731 } else if (nextPrefix.equals("#:") || nextPrefix.equals("##:")){
732 final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
733 lastForeign = nextLine;
734 if (pair.lang1 != "--" && pair.lang1 != "--") {
735 pairEntry.pairs.add(pair);
737 } else if (nextPrefix.equals("#::") || nextPrefix.equals("#**")) {
738 if (lastForeign != null && pairEntry.pairs.size() > 0) {
739 pairEntry.pairs.remove(pairEntry.pairs.size() - 1);
740 final Pair pair = new Pair(formatAndIndexExampleString(nextLine, enIndexBuilder, indexedEntry), formatAndIndexExampleString(lastForeign, foreignIndexBuilder, indexedEntry), swap);
741 if (pair.lang1 != "--" || pair.lang2 != "--") {
742 pairEntry.pairs.add(pair);
746 LOG.warning("TODO: English example with no foreign: " + title + ", " + nextLine);
747 final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
748 if (pair.lang1 != "--" || pair.lang2 != "--") {
749 pairEntry.pairs.add(pair);
752 } else if (nextPrefix.equals("#*")) {
753 // Can't really index these.
754 final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
755 lastForeign = nextLine;
756 if (pair.lang1 != "--" || pair.lang2 != "--") {
757 pairEntry.pairs.add(pair);
759 } else if (nextPrefix.equals("#::*") || nextPrefix.equals("##") || nextPrefix.equals("#*:") || nextPrefix.equals("#:*") || true) {
760 final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
761 if (pair.lang1 != "--" || pair.lang2 != "--") {
762 pairEntry.pairs.add(pair);
770 private String formatAndIndexExampleString(final String example, final IndexBuilder indexBuilder, final IndexedEntry indexedEntry) {
771 final WikiTokenizer wikiTokenizer = new WikiTokenizer(example, false);
772 final StringBuilder builder = new StringBuilder();
773 boolean insideTripleQuotes = false;
774 while (wikiTokenizer.nextToken() != null) {
775 if (wikiTokenizer.isPlainText()) {
776 builder.append(wikiTokenizer.token());
777 if (indexBuilder != null) {
778 indexBuilder.addEntryWithStringNoSingle(indexedEntry, wikiTokenizer.token(), EntryTypeName.WIKTIONARY_EXAMPLE);
780 } else if (wikiTokenizer.isWikiLink()) {
781 final String text = wikiTokenizer.wikiLinkText().replaceAll("'", "");
782 builder.append(text);
783 if (indexBuilder != null) {
784 indexBuilder.addEntryWithStringNoSingle(indexedEntry, text, EntryTypeName.WIKTIONARY_EXAMPLE);
786 } else if (wikiTokenizer.isFunction()) {
787 builder.append(wikiTokenizer.token());
788 } else if (wikiTokenizer.isMarkup()) {
789 if (wikiTokenizer.token().equals("'''")) {
790 insideTripleQuotes = !insideTripleQuotes;
792 } else if (wikiTokenizer.isComment() || wikiTokenizer.isNewline()) {
795 LOG.warning("unexpected token: " + wikiTokenizer.token());
798 final String result = trim(builder.toString());
799 return result.length() > 0 ? result : "--";
803 private void itConjAre(List<String> args, Map<String, String> namedArgs) {
804 final String base = args.get(0);
805 final String aux = args.get(1);
807 putIfMissing(namedArgs, "inf", base + "are");
808 putIfMissing(namedArgs, "aux", aux);
809 putIfMissing(namedArgs, "ger", base + "ando");
810 putIfMissing(namedArgs, "presp", base + "ante");
811 putIfMissing(namedArgs, "pastp", base + "ato");
813 putIfMissing(namedArgs, "pres1s", base + "o");
814 putIfMissing(namedArgs, "pres2s", base + "i");
815 putIfMissing(namedArgs, "pres3s", base + "a");
816 putIfMissing(namedArgs, "pres1p", base + "iamo");
817 putIfMissing(namedArgs, "pres2p", base + "ate");
818 putIfMissing(namedArgs, "pres3p", base + "ano");
820 putIfMissing(namedArgs, "imperf1s", base + "avo");
821 putIfMissing(namedArgs, "imperf2s", base + "avi");
822 putIfMissing(namedArgs, "imperf3s", base + "ava");
823 putIfMissing(namedArgs, "imperf1p", base + "avamo");
824 putIfMissing(namedArgs, "imperf2p", base + "avate");
825 putIfMissing(namedArgs, "imperf3p", base + "avano");
827 putIfMissing(namedArgs, "prem1s", base + "ai");
828 putIfMissing(namedArgs, "prem2s", base + "asti");
829 putIfMissing(namedArgs, "prem3s", base + "ò");
830 putIfMissing(namedArgs, "prem1p", base + "ammo");
831 putIfMissing(namedArgs, "prem2p", base + "aste");
832 putIfMissing(namedArgs, "prem3p", base + "arono");
834 putIfMissing(namedArgs, "fut1s", base + "erò");
835 putIfMissing(namedArgs, "fut2s", base + "erai");
836 putIfMissing(namedArgs, "fut3s", base + "erà");
837 putIfMissing(namedArgs, "fut1p", base + "eremo");
838 putIfMissing(namedArgs, "fut2p", base + "erete");
839 putIfMissing(namedArgs, "fut3p", base + "eranno");
841 putIfMissing(namedArgs, "cond1s", base + "erei");
842 putIfMissing(namedArgs, "cond2s", base + "eresti");
843 putIfMissing(namedArgs, "cond3s", base + "erebbe");
844 putIfMissing(namedArgs, "cond1p", base + "eremmo");
845 putIfMissing(namedArgs, "cond2p", base + "ereste");
846 putIfMissing(namedArgs, "cond3p", base + "erebbero");
847 // Subjunctive / congiuntivo
848 putIfMissing(namedArgs, "sub123s", base + "i");
849 putIfMissing(namedArgs, "sub1p", base + "iamo");
850 putIfMissing(namedArgs, "sub2p", base + "iate");
851 putIfMissing(namedArgs, "sub3p", base + "ino");
852 // Imperfect subjunctive
853 putIfMissing(namedArgs, "impsub12s", base + "assi");
854 putIfMissing(namedArgs, "impsub3s", base + "asse");
855 putIfMissing(namedArgs, "impsub1p", base + "assimo");
856 putIfMissing(namedArgs, "impsub2p", base + "aste");
857 putIfMissing(namedArgs, "impsub3p", base + "assero");
859 putIfMissing(namedArgs, "imp2s", base + "a");
860 putIfMissing(namedArgs, "imp3s", base + "i");
861 putIfMissing(namedArgs, "imp1p", base + "iamo");
862 putIfMissing(namedArgs, "imp2p", base + "ate");
863 putIfMissing(namedArgs, "imp3p", base + "ino");
866 itConj(args, namedArgs);
870 private void itConj(List<String> args, Map<String, String> namedArgs) {
871 // TODO Auto-generated method stub
876 private static void putIfMissing(final Map<String, String> namedArgs, final String key,
877 final String value) {
878 final String oldValue = namedArgs.get(key);
879 if (oldValue == null || oldValue.length() == 0) {
880 namedArgs.put(key, value);
884 // TODO: check how ='' and =| are manifested....
885 // TODO: get this right in -are
886 private static void putOrNullify(final Map<String, String> namedArgs, final String key,
887 final String value) {
888 final String oldValue = namedArgs.get(key);
889 if (oldValue == null/* || oldValue.length() == 0*/) {
890 namedArgs.put(key, value);
892 if (oldValue.equals("''")) {
893 namedArgs.put(key, "");
898 static final Pattern whitespace = Pattern.compile("\\s+");
899 static String trim(final String s) {
900 return whitespace.matcher(s).replaceAll(" ").trim();