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.Level;
31 import java.util.logging.Logger;
32 import java.util.regex.Pattern;
34 import com.hughes.android.dictionary.engine.EntrySource;
35 import com.hughes.android.dictionary.engine.EntryTypeName;
36 import com.hughes.android.dictionary.engine.IndexBuilder;
37 import com.hughes.android.dictionary.engine.IndexedEntry;
38 import com.hughes.android.dictionary.engine.PairEntry;
39 import com.hughes.android.dictionary.engine.PairEntry.Pair;
40 import com.hughes.android.dictionary.parser.WikiTokenizer;
42 public class EnWiktionaryXmlParser {
44 static final Logger LOG = Logger.getLogger(EnWiktionaryXmlParser.class.getName());
46 // TODO: process {{ttbc}} lines
48 static final Pattern partOfSpeechHeader = Pattern.compile(
49 "Noun|Verb|Adjective|Adverb|Pronoun|Conjunction|Interjection|" +
50 "Preposition|Proper noun|Article|Prepositional phrase|Acronym|" +
51 "Abbreviation|Initialism|Contraction|Prefix|Suffix|Symbol|Letter|" +
52 "Ligature|Idiom|Phrase|\\{\\{acronym\\}\\}|\\{\\{initialism\\}\\}|" +
53 "\\{\\{abbreviation\\}\\}|" +
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 EntrySource entrySource;
63 final IndexBuilder enIndexBuilder;
64 final IndexBuilder foreignIndexBuilder;
65 final Pattern langPattern;
66 final Pattern langCodePattern;
69 // State used while parsing.
72 ENGLISH_DEF_OF_FOREIGN,
79 public EnWiktionaryXmlParser(final IndexBuilder enIndexBuilder, final IndexBuilder otherIndexBuilder, final Pattern langPattern, final Pattern langCodePattern, final boolean swap) {
80 this.enIndexBuilder = enIndexBuilder;
81 this.foreignIndexBuilder = otherIndexBuilder;
82 this.langPattern = langPattern;
83 this.langCodePattern = langCodePattern;
88 public void parse(final File file, final EntrySource entrySource, final int pageLimit) throws IOException {
89 this.entrySource = entrySource;
91 final DataInputStream dis = new DataInputStream(new BufferedInputStream(new FileInputStream(file)));
93 if (pageLimit >= 0 && pageCount >= pageLimit) {
98 title = dis.readUTF();
99 } catch (EOFException e) {
100 LOG.log(Level.INFO, "EOF reading split.");
104 final String heading = dis.readUTF();
105 final int bytesLength = dis.readInt();
106 final byte[] bytes = new byte[bytesLength];
107 dis.readFully(bytes);
108 final String text = new String(bytes, "UTF8");
110 parseSection(heading, text);
113 if (pageCount % 1000 == 0) {
114 LOG.info("pageCount=" + pageCount);
119 private void parseSection(String heading, final String text) {
120 if (title.startsWith("Wiktionary:") ||
121 title.startsWith("Template:") ||
122 title.startsWith("Appendix:") ||
123 title.startsWith("Category:") ||
124 title.startsWith("Index:") ||
125 title.startsWith("MediaWiki:") ||
126 title.startsWith("TransWiki:") ||
127 title.startsWith("Citations:") ||
128 title.startsWith("Concordance:") ||
129 title.startsWith("Help:")) {
133 heading = heading.replaceAll("=", "").trim();
134 if (heading.equals("English")) {
136 } else if (langPattern.matcher(heading).find()){
137 doForeignWord(heading, text);
142 // -------------------------------------------------------------------------
144 private void doEnglishWord(String text) {
149 final WikiTokenizer wikiTokenizer = new WikiTokenizer(text);
150 while (wikiTokenizer.nextToken() != null) {
152 if (wikiTokenizer.isHeading()) {
153 final String headerName = wikiTokenizer.headingWikiText();
155 if (wikiTokenizer.headingDepth() <= posDepth) {
160 if (partOfSpeechHeader.matcher(headerName).matches()) {
161 posDepth = wikiTokenizer.headingDepth();
162 pos = wikiTokenizer.headingWikiText();
163 // TODO: if we're inside the POS section, we should handle the first title line...
165 } else if (headerName.equals("Translations")) {
167 LOG.info("Translations without POS (but using anyway): " + title);
169 doTranslations(wikiTokenizer, pos);
170 } else if (headerName.equals("Pronunciation")) {
171 //doPronunciation(wikiLineReader);
173 } else if (wikiTokenizer.isFunction()) {
174 final String name = wikiTokenizer.functionName();
175 if (name.equals("head") && pos == null) {
176 LOG.warning("{{head}} without POS: " + title);
182 final AppendAndIndexWikiCallback appendAndIndexWikiCallback = new AppendAndIndexWikiCallback(this);
184 appendAndIndexWikiCallback.functionCallbacks.putAll(FunctionCallbacksDefault.DEFAULT);
187 private void doTranslations(final WikiTokenizer wikiTokenizer, final String pos) {
188 if (title.equals("absolutely")) {
189 //System.out.println();
192 String topLevelLang = null;
194 boolean done = false;
195 while (wikiTokenizer.nextToken() != null) {
196 if (wikiTokenizer.isHeading()) {
197 wikiTokenizer.returnToLineStart();
204 // Check whether we care about this line:
206 if (wikiTokenizer.isFunction()) {
207 final String functionName = wikiTokenizer.functionName();
208 final List<String> positionArgs = wikiTokenizer.functionPositionArgs();
210 if (functionName.equals("trans-top")) {
212 if (wikiTokenizer.functionPositionArgs().size() >= 1) {
213 sense = positionArgs.get(0);
214 // TODO: could emphasize words in [[brackets]] inside sense.
215 sense = WikiTokenizer.toPlainText(sense);
216 //LOG.info("Sense: " + sense);
218 } else if (functionName.equals("trans-bottom")) {
220 } else if (functionName.equals("trans-mid")) {
221 } else if (functionName.equals("trans-see")) {
222 // TODO: would also be nice...
223 } else if (functionName.startsWith("picdic")) {
224 } else if (functionName.startsWith("checktrans")) {
226 } else if (functionName.startsWith("ttbc")) {
227 wikiTokenizer.nextLine();
228 // TODO: would be great to handle ttbc
229 // TODO: Check this: done = true;
231 LOG.warning("Unexpected translation wikifunction: " + wikiTokenizer.token() + ", title=" + title);
233 } else if (wikiTokenizer.isListItem()) {
234 final String line = wikiTokenizer.listItemWikiText();
235 // This line could produce an output...
237 if (line.contains("ich hoan dich gear")) {
238 //System.out.println();
241 // First strip the language and check whether it matches.
242 // And hold onto it for sub-lines.
243 final int colonIndex = line.indexOf(":");
244 if (colonIndex == -1) {
248 final String lang = trim(WikiTokenizer.toPlainText(line.substring(0, colonIndex)));
249 final boolean appendLang;
250 if (wikiTokenizer.listItemPrefix().length() == 1) {
252 final boolean thisFind = langPattern.matcher(lang).find();
256 appendLang = !langPattern.matcher(lang).matches();
257 } else if (topLevelLang == null) {
260 // Two-level -- the only way we won't append is if this second level matches exactly.
261 if (!langPattern.matcher(lang).matches() && !langPattern.matcher(topLevelLang).find()) {
264 appendLang = !langPattern.matcher(lang).matches();
267 String rest = line.substring(colonIndex + 1).trim();
268 if (rest.length() > 0) {
269 doTranslationLine(line, appendLang ? lang : null, pos, sense, rest);
272 } else if (wikiTokenizer.remainderStartsWith("''See''")) {
273 wikiTokenizer.nextLine();
274 LOG.fine("Skipping See line: " + wikiTokenizer.token());
275 } else if (wikiTokenizer.isWikiLink()) {
276 final String wikiLink = wikiTokenizer.wikiLinkText();
277 if (wikiLink.contains(":") && wikiLink.contains(title)) {
278 } else if (wikiLink.contains("Category:")) {
280 LOG.warning("Unexpected wikiLink: " + wikiTokenizer.token() + ", title=" + title);
282 } else if (wikiTokenizer.isNewline() || wikiTokenizer.isMarkup() || wikiTokenizer.isComment()) {
284 final String token = wikiTokenizer.token();
285 if (token.equals("----")) {
287 LOG.warning("Unexpected translation token: " + wikiTokenizer.token() + ", title=" + title);
294 private void doTranslationLine(final String line, final String lang, final String pos, final String sense, final String rest) {
295 state = State.TRANSLATION_LINE;
296 // Good chance we'll actually file this one...
297 final PairEntry pairEntry = new PairEntry(entrySource);
298 final IndexedEntry indexedEntry = new IndexedEntry(pairEntry);
300 final StringBuilder foreignText = new StringBuilder();
301 appendAndIndexWikiCallback.reset(foreignText, indexedEntry);
302 appendAndIndexWikiCallback.dispatch(rest, foreignIndexBuilder, EntryTypeName.WIKTIONARY_TRANSLATION_OTHER_TEXT);
304 if (foreignText.length() == 0) {
305 LOG.warning("Empty foreignText: " + line);
310 foreignText.insert(0, String.format("(%s) ", lang));
313 StringBuilder englishText = new StringBuilder();
315 englishText.append(title);
317 englishText.append(" (").append(sense).append(")");
318 enIndexBuilder.addEntryWithString(indexedEntry, sense, EntryTypeName.WIKTIONARY_TRANSLATION_SENSE);
321 englishText.append(" (").append(pos.toLowerCase()).append(")");
323 enIndexBuilder.addEntryWithString(indexedEntry, title, EntryTypeName.WIKTIONARY_TITLE_MULTI);
325 final Pair pair = new Pair(trim(englishText.toString()), trim(foreignText.toString()), swap);
326 pairEntry.pairs.add(pair);
327 if (!pairsAdded.add(pair.toString())) {
328 LOG.warning("Duplicate pair: " + pair.toString());
333 Set<String> pairsAdded = new LinkedHashSet<String>();
335 // -------------------------------------------------------------------------
337 private void doForeignWord(final String lang, final String text) {
338 final WikiTokenizer wikiTokenizer = new WikiTokenizer(text);
339 while (wikiTokenizer.nextToken() != null) {
340 if (wikiTokenizer.isHeading()) {
341 final String headingName = wikiTokenizer.headingWikiText();
342 if (headingName.equals("Translations")) {
343 LOG.warning("Translations not in English section: " + title);
344 } else if (headingName.equals("Pronunciation")) {
345 //doPronunciation(wikiLineReader);
346 } else if (partOfSpeechHeader.matcher(headingName).matches()) {
347 doForeignPartOfSpeech(lang, headingName, wikiTokenizer.headingDepth(), wikiTokenizer);
354 static final class ListSection {
355 final String firstPrefix;
356 final String firstLine;
357 final List<String> nextPrefixes = new ArrayList<String>();
358 final List<String> nextLines = new ArrayList<String>();
360 public ListSection(String firstPrefix, String firstLine) {
361 this.firstPrefix = firstPrefix;
362 this.firstLine = firstLine;
366 public String toString() {
367 return firstPrefix + firstLine + "{ " + nextPrefixes + "}";
372 int foreignCount = 0;
373 final Collection<String> wordForms = new ArrayList<String>();
374 boolean titleAppended = false;
376 private void doForeignPartOfSpeech(final String lang, String posHeading, final int posDepth, WikiTokenizer wikiTokenizer) {
377 if (++foreignCount % 1000 == 0) {
378 LOG.info("***" + lang + ", " + title + ", pos=" + posHeading + ", foreignCount=" + foreignCount);
380 if (title.equals("6")) {
381 System.out.println();
384 final StringBuilder foreignBuilder = new StringBuilder();
385 final List<ListSection> listSections = new ArrayList<ListSection>();
387 appendAndIndexWikiCallback.reset(foreignBuilder, null);
388 this.state = State.ENGLISH_DEF_OF_FOREIGN; // TODO: this is wrong, need new category....
389 titleAppended = false;
394 ListSection lastListSection = null;
396 int currentHeadingDepth = posDepth;
397 while (wikiTokenizer.nextToken() != null) {
398 if (wikiTokenizer.isHeading()) {
399 currentHeadingDepth = wikiTokenizer.headingDepth();
401 if (currentHeadingDepth <= posDepth) {
402 wikiTokenizer.returnToLineStart();
407 if (currentHeadingDepth > posDepth) {
408 // TODO: deal with other neat info sections
412 if (wikiTokenizer.isFunction()) {
413 final String name = wikiTokenizer.functionName();
414 final List<String> args = wikiTokenizer.functionPositionArgs();
415 final Map<String,String> namedArgs = wikiTokenizer.functionNamedArgs();
416 // First line is generally a repeat of the title with some extra information.
417 // We need to build up the left side (foreign text, tokens) separately from the
418 // right side (English). The left-side may get paired with multiple right sides.
419 // The left side should get filed under every form of the word in question (singular, plural).
421 // For verbs, the conjugation comes later on in a deeper section.
422 // Ideally, we'd want to file every English entry with the verb
423 // under every verb form coming from the conjugation.
424 // Ie. under "fa": see: "make :: fare" and "do :: fare"
425 // But then where should we put the conjugation table?
426 // 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!)
427 // for the conjugation table from "fa".
428 // Would like to be able to link to a lang#token.
430 appendAndIndexWikiCallback.onFunction(wikiTokenizer, name, args, namedArgs);
432 } else if (wikiTokenizer.isListItem()) {
433 final String prefix = wikiTokenizer.listItemPrefix();
434 if (lastListSection != null &&
435 prefix.startsWith(lastListSection.firstPrefix) &&
436 prefix.length() > lastListSection.firstPrefix.length()) {
437 lastListSection.nextPrefixes.add(prefix);
438 lastListSection.nextLines.add(wikiTokenizer.listItemWikiText());
440 lastListSection = new ListSection(prefix, wikiTokenizer.listItemWikiText());
441 listSections.add(lastListSection);
443 } else if (lastListSection != null) {
444 // Don't append anything after the lists, because there's crap.
445 } else if (wikiTokenizer.isWikiLink()) {
447 foreignBuilder.append(wikiTokenizer.wikiLinkText());
449 } else if (wikiTokenizer.isPlainText()) {
451 foreignBuilder.append(wikiTokenizer.token());
453 } else if (wikiTokenizer.isMarkup() || wikiTokenizer.isNewline() || wikiTokenizer.isComment()) {
456 LOG.warning("Unexpected token: " + wikiTokenizer.token());
461 // Here's where we exit.
462 // Should we make an entry even if there are no foreign list items?
463 String foreign = foreignBuilder.toString().trim();
464 if (!titleAppended && !foreign.toLowerCase().startsWith(title.toLowerCase())) {
465 foreign = String.format("%s %s", title, foreign);
467 if (!langPattern.matcher(lang).matches()) {
468 foreign = String.format("(%s) %s", lang, foreign);
470 for (final ListSection listSection : listSections) {
471 doForeignListSection(foreign, title, wordForms, listSection);
477 // Might only want to remove "lang" if it's equal to "zh", for example.
478 static final Set<String> USELESS_WIKI_ARGS = new LinkedHashSet<String>(
487 public boolean entryIsFormOfSomething = false;
489 private void doForeignListSection(final String foreignText, String title, final Collection<String> forms, final ListSection listSection) {
490 state = State.ENGLISH_DEF_OF_FOREIGN;
491 final String prefix = listSection.firstPrefix;
492 if (prefix.length() > 1) {
493 // Could just get looser and say that any prefix longer than first is a sublist.
494 LOG.warning("Prefix too long: " + listSection);
498 final PairEntry pairEntry = new PairEntry(entrySource);
499 final IndexedEntry indexedEntry = new IndexedEntry(pairEntry);
501 entryIsFormOfSomething = false;
502 final StringBuilder englishBuilder = new StringBuilder();
503 final String mainLine = listSection.firstLine;
504 appendAndIndexWikiCallback.reset(englishBuilder, indexedEntry);
505 appendAndIndexWikiCallback.dispatch(mainLine, enIndexBuilder, EntryTypeName.WIKTIONARY_ENGLISH_DEF);
507 final String english = trim(englishBuilder.toString());
508 if (english.length() > 0) {
509 final Pair pair = new Pair(english, trim(foreignText), this.swap);
510 pairEntry.pairs.add(pair);
511 foreignIndexBuilder.addEntryWithString(indexedEntry, title, entryIsFormOfSomething ? EntryTypeName.WIKTIONARY_IS_FORM_OF_SOMETHING_ELSE : EntryTypeName.WIKTIONARY_TITLE_MULTI);
512 for (final String form : forms) {
513 foreignIndexBuilder.addEntryWithString(indexedEntry, form, EntryTypeName.WIKTIONARY_INFLECTED_FORM_MULTI);
518 String lastForeign = null;
519 for (int i = 0; i < listSection.nextPrefixes.size(); ++i) {
520 final String nextPrefix = listSection.nextPrefixes.get(i);
521 final String nextLine = listSection.nextLines.get(i);
523 // TODO: This splitting is not sensitive to wiki code.
524 int dash = nextLine.indexOf("—");
527 dash = nextLine.indexOf("—");
531 dash = nextLine.indexOf(" - ");
535 if ((nextPrefix.equals("#:") || nextPrefix.equals("##:")) && dash != -1) {
536 final String foreignEx = nextLine.substring(0, dash);
537 final String englishEx = nextLine.substring(dash + mdashLen);
538 final Pair pair = new Pair(formatAndIndexExampleString(englishEx, enIndexBuilder, indexedEntry), formatAndIndexExampleString(foreignEx, foreignIndexBuilder, indexedEntry), swap);
539 if (pair.lang1 != "--" && pair.lang1 != "--") {
540 pairEntry.pairs.add(pair);
543 } else if (nextPrefix.equals("#:") || nextPrefix.equals("##:")){
544 final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
545 lastForeign = nextLine;
546 if (pair.lang1 != "--" && pair.lang1 != "--") {
547 pairEntry.pairs.add(pair);
549 } else if (nextPrefix.equals("#::") || nextPrefix.equals("#**")) {
550 if (lastForeign != null && pairEntry.pairs.size() > 0) {
551 pairEntry.pairs.remove(pairEntry.pairs.size() - 1);
552 final Pair pair = new Pair(formatAndIndexExampleString(nextLine, enIndexBuilder, indexedEntry), formatAndIndexExampleString(lastForeign, foreignIndexBuilder, indexedEntry), swap);
553 if (pair.lang1 != "--" || pair.lang2 != "--") {
554 pairEntry.pairs.add(pair);
558 LOG.warning("TODO: English example with no foreign: " + title + ", " + nextLine);
559 final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
560 if (pair.lang1 != "--" || pair.lang2 != "--") {
561 pairEntry.pairs.add(pair);
564 } else if (nextPrefix.equals("#*")) {
565 // Can't really index these.
566 final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
567 lastForeign = nextLine;
568 if (pair.lang1 != "--" || pair.lang2 != "--") {
569 pairEntry.pairs.add(pair);
571 } else if (nextPrefix.equals("#::*") || nextPrefix.equals("##") || nextPrefix.equals("#*:") || nextPrefix.equals("#:*") || true) {
572 final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
573 if (pair.lang1 != "--" || pair.lang2 != "--") {
574 pairEntry.pairs.add(pair);
582 private String formatAndIndexExampleString(final String example, final IndexBuilder indexBuilder, final IndexedEntry indexedEntry) {
584 // if (wikiTokenizer.token().equals("'''")) {
585 // insideTripleQuotes = !insideTripleQuotes;
587 final StringBuilder builder = new StringBuilder();
588 appendAndIndexWikiCallback.reset(builder, indexedEntry);
589 appendAndIndexWikiCallback.entryTypeName = EntryTypeName.WIKTIONARY_EXAMPLE;
590 appendAndIndexWikiCallback.entryTypeNameSticks = true;
592 // TODO: this is a hack needed because we don't safely split on the dash.
593 appendAndIndexWikiCallback.dispatch(example, indexBuilder, EntryTypeName.WIKTIONARY_EXAMPLE);
594 } catch (AssertionError e) {
597 final String result = trim(builder.toString());
598 return result.length() > 0 ? result : "--";
602 private void itConjAre(List<String> args, Map<String, String> namedArgs) {
603 final String base = args.get(0);
604 final String aux = args.get(1);
606 putIfMissing(namedArgs, "inf", base + "are");
607 putIfMissing(namedArgs, "aux", aux);
608 putIfMissing(namedArgs, "ger", base + "ando");
609 putIfMissing(namedArgs, "presp", base + "ante");
610 putIfMissing(namedArgs, "pastp", base + "ato");
612 putIfMissing(namedArgs, "pres1s", base + "o");
613 putIfMissing(namedArgs, "pres2s", base + "i");
614 putIfMissing(namedArgs, "pres3s", base + "a");
615 putIfMissing(namedArgs, "pres1p", base + "iamo");
616 putIfMissing(namedArgs, "pres2p", base + "ate");
617 putIfMissing(namedArgs, "pres3p", base + "ano");
619 putIfMissing(namedArgs, "imperf1s", base + "avo");
620 putIfMissing(namedArgs, "imperf2s", base + "avi");
621 putIfMissing(namedArgs, "imperf3s", base + "ava");
622 putIfMissing(namedArgs, "imperf1p", base + "avamo");
623 putIfMissing(namedArgs, "imperf2p", base + "avate");
624 putIfMissing(namedArgs, "imperf3p", base + "avano");
626 putIfMissing(namedArgs, "prem1s", base + "ai");
627 putIfMissing(namedArgs, "prem2s", base + "asti");
628 putIfMissing(namedArgs, "prem3s", base + "ò");
629 putIfMissing(namedArgs, "prem1p", base + "ammo");
630 putIfMissing(namedArgs, "prem2p", base + "aste");
631 putIfMissing(namedArgs, "prem3p", base + "arono");
633 putIfMissing(namedArgs, "fut1s", base + "erò");
634 putIfMissing(namedArgs, "fut2s", base + "erai");
635 putIfMissing(namedArgs, "fut3s", base + "erà");
636 putIfMissing(namedArgs, "fut1p", base + "eremo");
637 putIfMissing(namedArgs, "fut2p", base + "erete");
638 putIfMissing(namedArgs, "fut3p", base + "eranno");
640 putIfMissing(namedArgs, "cond1s", base + "erei");
641 putIfMissing(namedArgs, "cond2s", base + "eresti");
642 putIfMissing(namedArgs, "cond3s", base + "erebbe");
643 putIfMissing(namedArgs, "cond1p", base + "eremmo");
644 putIfMissing(namedArgs, "cond2p", base + "ereste");
645 putIfMissing(namedArgs, "cond3p", base + "erebbero");
646 // Subjunctive / congiuntivo
647 putIfMissing(namedArgs, "sub123s", base + "i");
648 putIfMissing(namedArgs, "sub1p", base + "iamo");
649 putIfMissing(namedArgs, "sub2p", base + "iate");
650 putIfMissing(namedArgs, "sub3p", base + "ino");
651 // Imperfect subjunctive
652 putIfMissing(namedArgs, "impsub12s", base + "assi");
653 putIfMissing(namedArgs, "impsub3s", base + "asse");
654 putIfMissing(namedArgs, "impsub1p", base + "assimo");
655 putIfMissing(namedArgs, "impsub2p", base + "aste");
656 putIfMissing(namedArgs, "impsub3p", base + "assero");
658 putIfMissing(namedArgs, "imp2s", base + "a");
659 putIfMissing(namedArgs, "imp3s", base + "i");
660 putIfMissing(namedArgs, "imp1p", base + "iamo");
661 putIfMissing(namedArgs, "imp2p", base + "ate");
662 putIfMissing(namedArgs, "imp3p", base + "ino");
665 itConj(args, namedArgs);
669 private void itConj(List<String> args, Map<String, String> namedArgs) {
670 // TODO Auto-generated method stub
675 private static void putIfMissing(final Map<String, String> namedArgs, final String key,
676 final String value) {
677 final String oldValue = namedArgs.get(key);
678 if (oldValue == null || oldValue.length() == 0) {
679 namedArgs.put(key, value);
683 // TODO: check how ='' and =| are manifested....
684 // TODO: get this right in -are
685 private static void putOrNullify(final Map<String, String> namedArgs, final String key,
686 final String value) {
687 final String oldValue = namedArgs.get(key);
688 if (oldValue == null/* || oldValue.length() == 0*/) {
689 namedArgs.put(key, value);
691 if (oldValue.equals("''")) {
692 namedArgs.put(key, "");
697 static final Pattern whitespace = Pattern.compile("\\s+");
698 static String trim(final String s) {
699 return whitespace.matcher(s).replaceAll(" ").trim();