1 // Copyright 2011 Google Inc. All Rights Reserved.
3 // Licensed under the Apache License, Version 2.0 (the "License");
4 // you may not use this file except in compliance with the License.
5 // You may obtain a copy of the License at
7 // http://www.apache.org/licenses/LICENSE-2.0
9 // Unless required by applicable law or agreed to in writing, software
10 // distributed under the License is distributed on an "AS IS" BASIS,
11 // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
12 // See the License for the specific language governing permissions and
13 // limitations under the License.
15 package com.hughes.android.dictionary.parser.enwiktionary;
17 import java.io.BufferedInputStream;
18 import java.io.DataInputStream;
19 import java.io.EOFException;
21 import java.io.FileInputStream;
22 import java.io.IOException;
23 import java.util.ArrayList;
24 import java.util.Arrays;
25 import java.util.Collection;
26 import java.util.LinkedHashSet;
27 import java.util.List;
30 import java.util.logging.Logger;
31 import java.util.regex.Pattern;
33 import com.hughes.android.dictionary.engine.EntryTypeName;
34 import com.hughes.android.dictionary.engine.IndexBuilder;
35 import com.hughes.android.dictionary.engine.IndexedEntry;
36 import com.hughes.android.dictionary.engine.PairEntry;
37 import com.hughes.android.dictionary.engine.PairEntry.Pair;
38 import com.hughes.android.dictionary.parser.WikiTokenizer;
40 public class EnWiktionaryXmlParser {
42 private static final String TRANSLITERATION_FORMAT = " (tr. %s)";
44 static final Logger LOG = Logger.getLogger(EnWiktionaryXmlParser.class.getName());
46 // TODO: process {{ttbc}} lines
48 static final Pattern partOfSpeechHeader = Pattern.compile(
49 "Noun|Verb|Adjective|Adverb|Pronoun|Conjunction|Interjection|" +
50 "Preposition|Proper noun|Article|Prepositional phrase|Acronym|" +
51 "Abbreviation|Initialism|Contraction|Prefix|Suffix|Symbol|Letter|" +
52 "Ligature|Idiom|Phrase|\\{\\{acronym\\}\\}|\\{\\{initialism\\}\\}|" +
53 "\\{\\{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 final IndexBuilder enIndexBuilder;
63 final IndexBuilder foreignIndexBuilder;
64 final Pattern langPattern;
65 final Pattern langCodePattern;
68 // State used while parsing.
71 ENGLISH_DEF_OF_FOREIGN,
78 public EnWiktionaryXmlParser(final IndexBuilder enIndexBuilder, final IndexBuilder otherIndexBuilder, final Pattern langPattern, final Pattern langCodePattern, final boolean swap) {
79 this.enIndexBuilder = enIndexBuilder;
80 this.foreignIndexBuilder = otherIndexBuilder;
81 this.langPattern = langPattern;
82 this.langCodePattern = langCodePattern;
87 public void parse(final File file, final int pageLimit) throws IOException {
89 final DataInputStream dis = new DataInputStream(new BufferedInputStream(new FileInputStream(file)));
91 if (pageLimit >= 0 && pageCount >= pageLimit) {
96 title = dis.readUTF();
97 } catch (EOFException e) {
98 LOG.warning("Error reading split!");
102 final String heading = dis.readUTF();
103 final int bytesLength = dis.readInt();
104 final byte[] bytes = new byte[bytesLength];
105 dis.readFully(bytes);
106 final String text = new String(bytes, "UTF8");
108 parseSection(heading, text);
111 if (pageCount % 1000 == 0) {
112 LOG.info("pageCount=" + pageCount);
117 private void parseSection(String heading, final String text) {
118 if (title.startsWith("Wiktionary:") ||
119 title.startsWith("Template:") ||
120 title.startsWith("Appendix:") ||
121 title.startsWith("Category:") ||
122 title.startsWith("Index:") ||
123 title.startsWith("MediaWiki:") ||
124 title.startsWith("TransWiki:") ||
125 title.startsWith("Citations:") ||
126 title.startsWith("Concordance:") ||
127 title.startsWith("Help:")) {
131 heading = heading.replaceAll("=", "").trim();
132 if (heading.equals("English")) {
134 } else if (langPattern.matcher(heading).find()){
135 doForeignWord(heading, text);
140 // -------------------------------------------------------------------------
142 private void doEnglishWord(String text) {
147 final WikiTokenizer wikiTokenizer = new WikiTokenizer(text);
148 while (wikiTokenizer.nextToken() != null) {
150 if (wikiTokenizer.isHeading()) {
151 final String headerName = wikiTokenizer.headingWikiText();
153 if (wikiTokenizer.headingDepth() <= posDepth) {
158 if (partOfSpeechHeader.matcher(headerName).matches()) {
159 posDepth = wikiTokenizer.headingDepth();
160 pos = wikiTokenizer.headingWikiText();
161 // TODO: if we're inside the POS section, we should handle the first title line...
163 } else if (headerName.equals("Translations")) {
165 LOG.info("Translations without POS (but using anyway): " + title);
167 doTranslations(wikiTokenizer, pos);
168 } else if (headerName.equals("Pronunciation")) {
169 //doPronunciation(wikiLineReader);
171 } else if (wikiTokenizer.isFunction()) {
172 final String name = wikiTokenizer.functionName();
173 if (name.equals("head")) {
174 LOG.warning("{{head}} without POS: " + title);
180 final AppendAndIndexWikiCallback appendAndIndexWikiCallback = new AppendAndIndexWikiCallback(this);
182 appendAndIndexWikiCallback.functionCallbacks.putAll(FunctionCallbacksDefault.DEFAULT);
185 private void doTranslations(final WikiTokenizer wikiTokenizer, final String pos) {
186 if (title.equals("absolutely")) {
187 //System.out.println();
190 String topLevelLang = null;
192 boolean done = false;
193 while (wikiTokenizer.nextToken() != null) {
194 if (wikiTokenizer.isHeading()) {
195 wikiTokenizer.returnToLineStart();
202 // Check whether we care about this line:
204 if (wikiTokenizer.isFunction()) {
205 final String functionName = wikiTokenizer.functionName();
206 final List<String> positionArgs = wikiTokenizer.functionPositionArgs();
208 if (functionName.equals("trans-top")) {
210 if (wikiTokenizer.functionPositionArgs().size() >= 1) {
211 sense = positionArgs.get(0);
212 // TODO: could emphasize words in [[brackets]] inside sense.
213 sense = WikiTokenizer.toPlainText(sense);
214 //LOG.info("Sense: " + sense);
216 } else if (functionName.equals("trans-bottom")) {
218 } else if (functionName.equals("trans-mid")) {
219 } else if (functionName.equals("trans-see")) {
220 // TODO: would also be nice...
221 } else if (functionName.startsWith("picdic")) {
222 } else if (functionName.startsWith("checktrans")) {
224 } else if (functionName.startsWith("ttbc")) {
225 wikiTokenizer.nextLine();
226 // TODO: would be great to handle ttbc
227 // TODO: Check this: done = true;
229 LOG.warning("Unexpected translation wikifunction: " + wikiTokenizer.token() + ", title=" + title);
231 } else if (wikiTokenizer.isListItem()) {
232 final String line = wikiTokenizer.listItemWikiText();
233 // This line could produce an output...
235 if (line.contains("ich hoan dich gear")) {
236 //System.out.println();
239 // First strip the language and check whether it matches.
240 // And hold onto it for sub-lines.
241 final int colonIndex = line.indexOf(":");
242 if (colonIndex == -1) {
246 final String lang = trim(WikiTokenizer.toPlainText(line.substring(0, colonIndex)));
247 final boolean appendLang;
248 if (wikiTokenizer.listItemPrefix().length() == 1) {
250 final boolean thisFind = langPattern.matcher(lang).find();
254 appendLang = !langPattern.matcher(lang).matches();
255 } else if (topLevelLang == null) {
258 // Two-level -- the only way we won't append is if this second level matches exactly.
259 if (!langPattern.matcher(lang).matches() && !langPattern.matcher(topLevelLang).find()) {
262 appendLang = !langPattern.matcher(lang).matches();
265 String rest = line.substring(colonIndex + 1).trim();
266 if (rest.length() > 0) {
267 doTranslationLine(line, appendLang ? lang : null, pos, sense, rest);
270 } else if (wikiTokenizer.remainderStartsWith("''See''")) {
271 wikiTokenizer.nextLine();
272 LOG.fine("Skipping See line: " + wikiTokenizer.token());
273 } else if (wikiTokenizer.isWikiLink()) {
274 final String wikiLink = wikiTokenizer.wikiLinkText();
275 if (wikiLink.contains(":") && wikiLink.contains(title)) {
276 } else if (wikiLink.contains("Category:")) {
278 LOG.warning("Unexpected wikiLink: " + wikiTokenizer.token() + ", title=" + title);
280 } else if (wikiTokenizer.isNewline() || wikiTokenizer.isMarkup() || wikiTokenizer.isComment()) {
282 final String token = wikiTokenizer.token();
283 if (token.equals("----")) {
285 LOG.warning("Unexpected translation token: " + wikiTokenizer.token() + ", title=" + title);
292 private void doTranslationLine(final String line, final String lang, final String pos, final String sense, final String rest) {
293 state = State.TRANSLATION_LINE;
294 // Good chance we'll actually file this one...
295 final PairEntry pairEntry = new PairEntry();
296 final IndexedEntry indexedEntry = new IndexedEntry(pairEntry);
298 final StringBuilder foreignText = new StringBuilder();
299 appendAndIndexWikiCallback.reset(foreignText, indexedEntry);
300 appendAndIndexWikiCallback.dispatch(rest, foreignIndexBuilder, EntryTypeName.WIKTIONARY_TRANSLATION_OTHER_TEXT);
302 if (foreignText.length() == 0) {
303 LOG.warning("Empty foreignText: " + line);
308 foreignText.insert(0, String.format("(%s) ", lang));
311 StringBuilder englishText = new StringBuilder();
313 englishText.append(title);
315 englishText.append(" (").append(sense).append(")");
316 enIndexBuilder.addEntryWithString(indexedEntry, sense, EntryTypeName.WIKTIONARY_TRANSLATION_SENSE);
319 englishText.append(" (").append(pos.toLowerCase()).append(")");
321 enIndexBuilder.addEntryWithString(indexedEntry, title, EntryTypeName.WIKTIONARY_TITLE_MULTI);
323 final Pair pair = new Pair(trim(englishText.toString()), trim(foreignText.toString()), swap);
324 pairEntry.pairs.add(pair);
325 if (!pairsAdded.add(pair.toString())) {
326 LOG.warning("Duplicate pair: " + pair.toString());
328 if (pair.toString().equals("libero {m} :: free (adjective)")) {
329 System.out.println();
335 Set<String> pairsAdded = new LinkedHashSet<String>();
337 // -------------------------------------------------------------------------
339 private void doForeignWord(final String lang, final String text) {
340 final WikiTokenizer wikiTokenizer = new WikiTokenizer(text);
341 while (wikiTokenizer.nextToken() != null) {
342 if (wikiTokenizer.isHeading()) {
343 final String headingName = wikiTokenizer.headingWikiText();
344 if (headingName.equals("Translations")) {
345 LOG.warning("Translations not in English section: " + title);
346 } else if (headingName.equals("Pronunciation")) {
347 //doPronunciation(wikiLineReader);
348 } else if (partOfSpeechHeader.matcher(headingName).matches()) {
349 doForeignPartOfSpeech(lang, headingName, wikiTokenizer.headingDepth(), wikiTokenizer);
356 static final class ListSection {
357 final String firstPrefix;
358 final String firstLine;
359 final List<String> nextPrefixes = new ArrayList<String>();
360 final List<String> nextLines = new ArrayList<String>();
362 public ListSection(String firstPrefix, String firstLine) {
363 this.firstPrefix = firstPrefix;
364 this.firstLine = firstLine;
368 public String toString() {
369 return firstPrefix + firstLine + "{ " + nextPrefixes + "}";
374 int foreignCount = 0;
375 final Collection<String> wordForms = new ArrayList<String>();
376 boolean titleAppended = false;
379 private void doForeignPartOfSpeech(final String lang, String posHeading, final int posDepth, WikiTokenizer wikiTokenizer) {
380 if (++foreignCount % 1000 == 0) {
381 LOG.info("***" + lang + ", " + title + ", pos=" + posHeading + ", foreignCount=" + foreignCount);
383 if (title.equals("6")) {
384 System.out.println();
387 final StringBuilder foreignBuilder = new StringBuilder();
388 final List<ListSection> listSections = new ArrayList<ListSection>();
390 appendAndIndexWikiCallback.reset(foreignBuilder, null);
391 this.state = State.ENGLISH_DEF_OF_FOREIGN; // TODO: this is wrong, need new category....
392 titleAppended = false;
397 ListSection lastListSection = null;
399 int currentHeadingDepth = posDepth;
400 while (wikiTokenizer.nextToken() != null) {
401 if (wikiTokenizer.isHeading()) {
402 currentHeadingDepth = wikiTokenizer.headingDepth();
404 if (currentHeadingDepth <= posDepth) {
405 wikiTokenizer.returnToLineStart();
410 if (currentHeadingDepth > posDepth) {
411 // TODO: deal with other neat info sections
415 if (wikiTokenizer.isFunction()) {
416 final String name = wikiTokenizer.functionName();
417 final List<String> args = wikiTokenizer.functionPositionArgs();
418 final Map<String,String> namedArgs = wikiTokenizer.functionNamedArgs();
419 // First line is generally a repeat of the title with some extra information.
420 // We need to build up the left side (foreign text, tokens) separately from the
421 // right side (English). The left-side may get paired with multiple right sides.
422 // The left side should get filed under every form of the word in question (singular, plural).
424 // For verbs, the conjugation comes later on in a deeper section.
425 // Ideally, we'd want to file every English entry with the verb
426 // under every verb form coming from the conjugation.
427 // Ie. under "fa": see: "make :: fare" and "do :: fare"
428 // But then where should we put the conjugation table?
429 // 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!)
430 // for the conjugation table from "fa".
431 // Would like to be able to link to a lang#token.
433 appendAndIndexWikiCallback.onFunction(wikiTokenizer, name, args, namedArgs);
435 } else if (wikiTokenizer.isListItem()) {
436 final String prefix = wikiTokenizer.listItemPrefix();
437 if (lastListSection != null &&
438 prefix.startsWith(lastListSection.firstPrefix) &&
439 prefix.length() > lastListSection.firstPrefix.length()) {
440 lastListSection.nextPrefixes.add(prefix);
441 lastListSection.nextLines.add(wikiTokenizer.listItemWikiText());
443 lastListSection = new ListSection(prefix, wikiTokenizer.listItemWikiText());
444 listSections.add(lastListSection);
446 } else if (lastListSection != null) {
447 // Don't append anything after the lists, because there's crap.
448 } else if (wikiTokenizer.isWikiLink()) {
450 foreignBuilder.append(wikiTokenizer.wikiLinkText());
452 } else if (wikiTokenizer.isPlainText()) {
454 foreignBuilder.append(wikiTokenizer.token());
456 } else if (wikiTokenizer.isMarkup() || wikiTokenizer.isNewline() || wikiTokenizer.isComment()) {
459 LOG.warning("Unexpected token: " + wikiTokenizer.token());
464 // Here's where we exit.
465 // Should we make an entry even if there are no foreign list items?
466 String foreign = foreignBuilder.toString().trim();
467 if (!titleAppended && !foreign.toLowerCase().startsWith(title.toLowerCase())) {
468 foreign = String.format("%s %s", title, foreign);
470 if (!langPattern.matcher(lang).matches()) {
471 foreign = String.format("(%s) %s", lang, foreign);
473 for (final ListSection listSection : listSections) {
474 doForeignListItem(foreign, title, wordForms, listSection);
480 // Might only want to remove "lang" if it's equal to "zh", for example.
481 static final Set<String> USELESS_WIKI_ARGS = new LinkedHashSet<String>(
489 private void doForeignListItem(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();
499 final IndexedEntry indexedEntry = new IndexedEntry(pairEntry);
501 final StringBuilder englishBuilder = new StringBuilder();
502 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, 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();