1 // Copyright 2012 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.wiktionary;
17 import java.util.ArrayList;
18 import java.util.Collection;
19 import java.util.List;
21 import java.util.regex.Pattern;
23 import com.hughes.android.dictionary.engine.EntryTypeName;
24 import com.hughes.android.dictionary.engine.IndexBuilder;
25 import com.hughes.android.dictionary.engine.IndexedEntry;
26 import com.hughes.android.dictionary.engine.PairEntry;
27 import com.hughes.android.dictionary.engine.PairEntry.Pair;
28 import com.hughes.android.dictionary.parser.WikiTokenizer;
30 public final class EnForeignParser extends EnParser {
32 public EnForeignParser(final IndexBuilder enIndexBuilder,
33 final IndexBuilder otherIndexBuilder, final Pattern langPattern,
34 final Pattern langCodePattern, final boolean swap) {
35 super(enIndexBuilder, otherIndexBuilder, langPattern, langCodePattern, swap);
39 void parseSection(String heading, String text) {
40 if (isIgnorableTitle(title)) {
43 final String lang = heading.replaceAll("=", "").trim();
44 if (!langPattern.matcher(lang).find()){
48 final WikiTokenizer wikiTokenizer = new WikiTokenizer(text);
49 while (wikiTokenizer.nextToken() != null) {
50 if (wikiTokenizer.isHeading()) {
51 final String headingName = wikiTokenizer.headingWikiText();
52 if (headingName.equals("Translations")) {
53 LOG.warning("Translations not in English section: " + title);
54 incrementCount("WARNING: Translations not in English section");
55 } else if (headingName.equals("Pronunciation")) {
56 //doPronunciation(wikiLineReader);
57 } else if (partOfSpeechHeader.matcher(headingName).matches()) {
58 doForeignPartOfSpeech(lang, headingName, wikiTokenizer.headingDepth(), wikiTokenizer);
61 // It's not a heading.
62 // TODO: optimization: skip to next heading.
67 static final class ListSection {
68 final String firstPrefix;
69 final String firstLine;
70 final List<String> nextPrefixes = new ArrayList<String>();
71 final List<String> nextLines = new ArrayList<String>();
73 public ListSection(String firstPrefix, String firstLine) {
74 this.firstPrefix = firstPrefix;
75 this.firstLine = firstLine;
79 public String toString() {
80 return firstPrefix + firstLine + "{ " + nextPrefixes + "}";
85 private void doForeignPartOfSpeech(final String lang, String posHeading, final int posDepth, WikiTokenizer wikiTokenizer) {
86 if (++foreignCount % 1000 == 0) {
87 LOG.info("***" + lang + ", " + title + ", pos=" + posHeading + ", foreignCount=" + foreignCount);
89 if (title.equals("6")) {
93 final StringBuilder foreignBuilder = new StringBuilder();
94 final List<EnForeignParser.ListSection> listSections = new ArrayList<EnForeignParser.ListSection>();
96 appendAndIndexWikiCallback.reset(foreignBuilder, null);
97 this.state = State.ENGLISH_DEF_OF_FOREIGN; // TODO: this is wrong, need new category....
98 titleAppended = false;
103 EnForeignParser.ListSection lastListSection = null;
105 int currentHeadingDepth = posDepth;
106 while (wikiTokenizer.nextToken() != null) {
107 if (wikiTokenizer.isHeading()) {
108 currentHeadingDepth = wikiTokenizer.headingDepth();
110 if (currentHeadingDepth <= posDepth) {
111 wikiTokenizer.returnToLineStart();
116 if (currentHeadingDepth > posDepth) {
117 // TODO: deal with other neat info sections inside POS
121 if (wikiTokenizer.isFunction()) {
122 final String name = wikiTokenizer.functionName();
123 final List<String> args = wikiTokenizer.functionPositionArgs();
124 final Map<String,String> namedArgs = wikiTokenizer.functionNamedArgs();
125 // First line is generally a repeat of the title with some extra information.
126 // We need to build up the left side (foreign text, tokens) separately from the
127 // right side (English). The left-side may get paired with multiple right sides.
128 // The left side should get filed under every form of the word in question (singular, plural).
130 // For verbs, the conjugation comes later on in a deeper section.
131 // Ideally, we'd want to file every English entry with the verb
132 // under every verb form coming from the conjugation.
133 // Ie. under "fa": see: "make :: fare" and "do :: fare"
134 // But then where should we put the conjugation table?
135 // 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!)
136 // for the conjugation table from "fa".
137 // Would like to be able to link to a lang#token.
139 appendAndIndexWikiCallback.onFunction(wikiTokenizer, name, args, namedArgs);
141 } else if (wikiTokenizer.isListItem()) {
142 final String prefix = wikiTokenizer.listItemPrefix();
143 if (lastListSection != null &&
144 prefix.startsWith(lastListSection.firstPrefix) &&
145 prefix.length() > lastListSection.firstPrefix.length()) {
146 lastListSection.nextPrefixes.add(prefix);
147 lastListSection.nextLines.add(wikiTokenizer.listItemWikiText());
149 lastListSection = new ListSection(prefix, wikiTokenizer.listItemWikiText());
150 listSections.add(lastListSection);
152 } else if (lastListSection != null) {
153 // Don't append anything after the lists, because there's crap.
154 } else if (wikiTokenizer.isWikiLink()) {
156 foreignBuilder.append(wikiTokenizer.wikiLinkText());
158 } else if (wikiTokenizer.isPlainText()) {
160 foreignBuilder.append(wikiTokenizer.token());
162 } else if (wikiTokenizer.isMarkup() || wikiTokenizer.isNewline() || wikiTokenizer.isComment()) {
165 LOG.warning("Unexpected token: " + wikiTokenizer.token());
166 assert !wikiTokenizer.errors().isEmpty();
171 // Here's where we exit.
172 // Should we make an entry even if there are no foreign list items?
173 String foreign = foreignBuilder.toString().trim();
174 if (!titleAppended && !foreign.toLowerCase().startsWith(title.toLowerCase())) {
175 foreign = String.format("%s %s", title, foreign);
177 if (!langPattern.matcher(lang).matches()) {
178 foreign = String.format("(%s) %s", lang, foreign);
180 for (final EnForeignParser.ListSection listSection : listSections) {
181 doForeignListSection(foreign, title, wordForms, listSection);
186 private void doForeignListSection(final String foreignText, String title, final Collection<String> forms, final EnForeignParser.ListSection listSection) {
187 state = State.ENGLISH_DEF_OF_FOREIGN;
188 final String prefix = listSection.firstPrefix;
189 if (prefix.length() > 1) {
190 // Could just get looser and say that any prefix longer than first is a sublist.
191 LOG.warning("Prefix too long: " + listSection);
192 incrementCount("WARNING: Prefix too long");
196 final PairEntry pairEntry = new PairEntry(entrySource);
197 final IndexedEntry indexedEntry = new IndexedEntry(pairEntry);
199 entryIsFormOfSomething = false;
200 final StringBuilder englishBuilder = new StringBuilder();
201 final String mainLine = listSection.firstLine;
202 appendAndIndexWikiCallback.reset(englishBuilder, indexedEntry);
203 appendAndIndexWikiCallback.dispatch(mainLine, enIndexBuilder, EntryTypeName.WIKTIONARY_ENGLISH_DEF);
205 final String english = trim(englishBuilder.toString());
206 if (english.length() > 0) {
207 final Pair pair = new Pair(english, trim(foreignText), this.swap);
208 pairEntry.pairs.add(pair);
209 foreignIndexBuilder.addEntryWithString(indexedEntry, title, entryIsFormOfSomething ? EntryTypeName.WIKTIONARY_IS_FORM_OF_SOMETHING_ELSE : EntryTypeName.WIKTIONARY_TITLE_MULTI);
210 for (final String form : forms) {
211 foreignIndexBuilder.addEntryWithString(indexedEntry, form, EntryTypeName.WIKTIONARY_INFLECTED_FORM_MULTI);
216 String lastForeign = null;
217 for (int i = 0; i < listSection.nextPrefixes.size(); ++i) {
218 final String nextPrefix = listSection.nextPrefixes.get(i);
219 final String nextLine = listSection.nextLines.get(i);
221 // TODO: This splitting is not sensitive to wiki code.
222 int dash = nextLine.indexOf("—");
225 dash = nextLine.indexOf("—");
229 dash = nextLine.indexOf(" - ");
233 if ((nextPrefix.equals("#:") || nextPrefix.equals("##:")) && dash != -1) {
234 final String foreignEx = nextLine.substring(0, dash);
235 final String englishEx = nextLine.substring(dash + mdashLen);
236 final Pair pair = new Pair(formatAndIndexExampleString(englishEx, enIndexBuilder, indexedEntry), formatAndIndexExampleString(foreignEx, foreignIndexBuilder, indexedEntry), swap);
237 if (pair.lang1 != "--" && pair.lang1 != "--") {
238 pairEntry.pairs.add(pair);
241 } else if (nextPrefix.equals("#:") || nextPrefix.equals("##:")){
242 final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
243 lastForeign = nextLine;
244 if (pair.lang1 != "--" && pair.lang1 != "--") {
245 pairEntry.pairs.add(pair);
247 } else if (nextPrefix.equals("#::") || nextPrefix.equals("#**")) {
248 if (lastForeign != null && pairEntry.pairs.size() > 0) {
249 pairEntry.pairs.remove(pairEntry.pairs.size() - 1);
250 final Pair pair = new Pair(formatAndIndexExampleString(nextLine, enIndexBuilder, indexedEntry), formatAndIndexExampleString(lastForeign, foreignIndexBuilder, indexedEntry), swap);
251 if (pair.lang1 != "--" || pair.lang2 != "--") {
252 pairEntry.pairs.add(pair);
256 LOG.warning("TODO: English example with no foreign: " + title + ", " + nextLine);
257 final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
258 if (pair.lang1 != "--" || pair.lang2 != "--") {
259 pairEntry.pairs.add(pair);
262 } else if (nextPrefix.equals("#*")) {
263 // Can't really index these.
264 final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
265 lastForeign = nextLine;
266 if (pair.lang1 != "--" || pair.lang2 != "--") {
267 pairEntry.pairs.add(pair);
269 } else if (nextPrefix.equals("#::*") || nextPrefix.equals("##") || nextPrefix.equals("#*:") || nextPrefix.equals("#:*") || true) {
270 final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
271 if (pair.lang1 != "--" || pair.lang2 != "--") {
272 pairEntry.pairs.add(pair);
280 private String formatAndIndexExampleString(final String example, final IndexBuilder indexBuilder, final IndexedEntry indexedEntry) {
282 // if (wikiTokenizer.token().equals("'''")) {
283 // insideTripleQuotes = !insideTripleQuotes;
285 final StringBuilder builder = new StringBuilder();
286 appendAndIndexWikiCallback.reset(builder, indexedEntry);
287 appendAndIndexWikiCallback.entryTypeName = EntryTypeName.WIKTIONARY_EXAMPLE;
288 appendAndIndexWikiCallback.entryTypeNameSticks = true;
290 // TODO: this is a hack needed because we don't safely split on the dash.
291 appendAndIndexWikiCallback.dispatch(example, indexBuilder, EntryTypeName.WIKTIONARY_EXAMPLE);
292 } catch (AssertionError e) {
295 final String result = trim(builder.toString());
296 return result.length() > 0 ? result : "--";
300 private void itConjAre(List<String> args, Map<String, String> namedArgs) {
301 final String base = args.get(0);
302 final String aux = args.get(1);
304 putIfMissing(namedArgs, "inf", base + "are");
305 putIfMissing(namedArgs, "aux", aux);
306 putIfMissing(namedArgs, "ger", base + "ando");
307 putIfMissing(namedArgs, "presp", base + "ante");
308 putIfMissing(namedArgs, "pastp", base + "ato");
310 putIfMissing(namedArgs, "pres1s", base + "o");
311 putIfMissing(namedArgs, "pres2s", base + "i");
312 putIfMissing(namedArgs, "pres3s", base + "a");
313 putIfMissing(namedArgs, "pres1p", base + "iamo");
314 putIfMissing(namedArgs, "pres2p", base + "ate");
315 putIfMissing(namedArgs, "pres3p", base + "ano");
317 putIfMissing(namedArgs, "imperf1s", base + "avo");
318 putIfMissing(namedArgs, "imperf2s", base + "avi");
319 putIfMissing(namedArgs, "imperf3s", base + "ava");
320 putIfMissing(namedArgs, "imperf1p", base + "avamo");
321 putIfMissing(namedArgs, "imperf2p", base + "avate");
322 putIfMissing(namedArgs, "imperf3p", base + "avano");
324 putIfMissing(namedArgs, "prem1s", base + "ai");
325 putIfMissing(namedArgs, "prem2s", base + "asti");
326 putIfMissing(namedArgs, "prem3s", base + "ò");
327 putIfMissing(namedArgs, "prem1p", base + "ammo");
328 putIfMissing(namedArgs, "prem2p", base + "aste");
329 putIfMissing(namedArgs, "prem3p", base + "arono");
331 putIfMissing(namedArgs, "fut1s", base + "erò");
332 putIfMissing(namedArgs, "fut2s", base + "erai");
333 putIfMissing(namedArgs, "fut3s", base + "erà");
334 putIfMissing(namedArgs, "fut1p", base + "eremo");
335 putIfMissing(namedArgs, "fut2p", base + "erete");
336 putIfMissing(namedArgs, "fut3p", base + "eranno");
338 putIfMissing(namedArgs, "cond1s", base + "erei");
339 putIfMissing(namedArgs, "cond2s", base + "eresti");
340 putIfMissing(namedArgs, "cond3s", base + "erebbe");
341 putIfMissing(namedArgs, "cond1p", base + "eremmo");
342 putIfMissing(namedArgs, "cond2p", base + "ereste");
343 putIfMissing(namedArgs, "cond3p", base + "erebbero");
344 // Subjunctive / congiuntivo
345 putIfMissing(namedArgs, "sub123s", base + "i");
346 putIfMissing(namedArgs, "sub1p", base + "iamo");
347 putIfMissing(namedArgs, "sub2p", base + "iate");
348 putIfMissing(namedArgs, "sub3p", base + "ino");
349 // Imperfect subjunctive
350 putIfMissing(namedArgs, "impsub12s", base + "assi");
351 putIfMissing(namedArgs, "impsub3s", base + "asse");
352 putIfMissing(namedArgs, "impsub1p", base + "assimo");
353 putIfMissing(namedArgs, "impsub2p", base + "aste");
354 putIfMissing(namedArgs, "impsub3p", base + "assero");
356 putIfMissing(namedArgs, "imp2s", base + "a");
357 putIfMissing(namedArgs, "imp3s", base + "i");
358 putIfMissing(namedArgs, "imp1p", base + "iamo");
359 putIfMissing(namedArgs, "imp2p", base + "ate");
360 putIfMissing(namedArgs, "imp3p", base + "ino");
363 itConj(args, namedArgs);
367 private void itConj(List<String> args, Map<String, String> namedArgs) {
368 // TODO Auto-generated method stub
373 private static void putIfMissing(final Map<String, String> namedArgs, final String key,
374 final String value) {
375 final String oldValue = namedArgs.get(key);
376 if (oldValue == null || oldValue.length() == 0) {
377 namedArgs.put(key, value);
381 // TODO: check how ='' and =| are manifested....
382 // TODO: get this right in -are
383 private static void putOrNullify(final Map<String, String> namedArgs, final String key,
384 final String value) {
385 final String oldValue = namedArgs.get(key);
386 if (oldValue == null/* || oldValue.length() == 0*/) {
387 namedArgs.put(key, value);
389 if (oldValue.equals("''")) {
390 namedArgs.put(key, "");