public class EnWiktionaryXmlParser {
+ private static final String TRANSLITERATION_FORMAT = " (tr. %s)";
+
static final Logger LOG = Logger.getLogger(EnWiktionaryXmlParser.class.getName());
- // TODO: look for {{ and [[ and <adf> <!-- in output.
// TODO: process {{ttbc}} lines
static final Pattern partOfSpeechHeader = Pattern.compile(
"Noun|Verb|Adjective|Adverb|Pronoun|Conjunction|Interjection|" +
"Preposition|Proper noun|Article|Prepositional phrase|Acronym|" +
"Abbreviation|Initialism|Contraction|Prefix|Suffix|Symbol|Letter|" +
- "Ligature|Idiom|Phrase|" +
+ "Ligature|Idiom|Phrase|\\{\\{acronym\\}\\}|\\{\\{initialism\\}\\}|" +
// These are @deprecated:
"Noun form|Verb form|Adjective form|Nominal phrase|Noun phrase|" +
"Verb phrase|Transitive verb|Intransitive verb|Reflexive verb|" +
heading = heading.replaceAll("=", "").trim();
if (heading.equals("English")) {
doEnglishWord(title, text);
- } else if (langPattern.matcher(heading).matches()){
- doForeignWord(title, text);
+ } else if (langPattern.matcher(heading).find()){
+ doForeignWord(heading, title, text);
}
} // endPage()
if (partOfSpeechHeader.matcher(headerName).matches()) {
posDepth = wikiTokenizer.headingDepth();
pos = wikiTokenizer.headingWikiText();
+ // TODO: if we're inside the POS section, we should handle the first title line...
+
} else if (headerName.equals("Translations")) {
if (pos == null) {
LOG.warning("Translations without POS: " + title);
private void doTranslations(final String title, final WikiTokenizer wikiTokenizer, final String pos) {
if (title.equals("absolutely")) {
- System.out.println();
+ //System.out.println();
}
+ String topLevelLang = null;
String sense = null;
boolean done = false;
while (wikiTokenizer.nextToken() != null) {
// Check whether we care about this line:
- //line = WikiLineReader.removeSquareBrackets(line);
-
if (wikiTokenizer.isFunction()) {
final String functionName = wikiTokenizer.functionName();
final List<String> positionArgs = wikiTokenizer.functionPositionArgs();
sense = null;
} else if (functionName.equals("trans-mid")) {
} else if (functionName.equals("trans-see")) {
- // TODO
+ // TODO: would also be nice...
} else if (functionName.startsWith("picdic")) {
} else if (functionName.startsWith("checktrans")) {
+ done = true;
} else if (functionName.startsWith("ttbc")) {
wikiTokenizer.nextLine();
- // TODO: would be great to handle
- //TODO: Check this: done = true;
+ // TODO: would be great to handle ttbc
+ // TODO: Check this: done = true;
} else {
LOG.warning("Unexpected translation wikifunction: " + wikiTokenizer.token() + ", title=" + title);
}
final String line = wikiTokenizer.listItemWikiText();
// This line could produce an output...
+ if (line.contains("ich hoan dich gear")) {
+ //System.out.println();
+ }
+
// First strip the language and check whether it matches.
// And hold onto it for sub-lines.
final int colonIndex = line.indexOf(":");
continue;
}
- final String lang = line.substring(0, colonIndex);
- if (!this.langPattern.matcher(lang).find()) {
+ final String lang = trim(WikiTokenizer.toPlainText(line.substring(0, colonIndex)));
+ final boolean appendLang;
+ if (wikiTokenizer.listItemPrefix().length() == 1) {
+ topLevelLang = lang;
+ final boolean thisFind = langPattern.matcher(lang).find();
+ if (!thisFind) {
+ continue;
+ }
+ appendLang = !langPattern.matcher(lang).matches();
+ } else if (topLevelLang == null) {
continue;
+ } else {
+ // Two-level -- the only way we won't append is if this second level matches exactly.
+ if (!langPattern.matcher(lang).matches() && !langPattern.matcher(topLevelLang).find()) {
+ continue;
+ }
+ appendLang = !langPattern.matcher(lang).matches();
}
String rest = line.substring(colonIndex + 1).trim();
if (rest.length() > 0) {
- doTranslationLine(line, title, pos, sense, rest);
- } else {
- // TODO: do lines that are like Greek:
+ doTranslationLine(line, appendLang ? lang : null, title, pos, sense, rest);
}
} else if (wikiTokenizer.remainderStartsWith("''See''")) {
}
}
+ private static <T> T get(final List<T> list, final int index, final T defaultValue) {
+ return index < list.size() ? list.get(index) : defaultValue;
+ }
+
private static <T> T get(final List<T> list, final int index) {
- return index < list.size() ? list.get(index) : null;
+ return get(list, index, null);
}
-
- private void doTranslationLine(final String line, final String title, final String pos, final String sense, final String rest) {
+
+ private static <T> T remove(final List<T> list, final int index, final T defaultValue) {
+ return index < list.size() ? list.remove(index) : defaultValue;
+ }
+
+ private void doTranslationLine(final String line, final String lang, final String title, final String pos, final String sense, final String rest) {
// Good chance we'll actually file this one...
final PairEntry pairEntry = new PairEntry();
final IndexedEntry indexedEntry = new IndexedEntry(pairEntry);
otherText.append(String.format(" {%s}", gender));
}
if (transliteration != null) {
- otherText.append(String.format(" (tr. %s)", transliteration));
+ otherText.append(String.format(TRANSLITERATION_FORMAT, transliteration));
otherIndexBuilder.addEntryWithString(indexedEntry, transliteration, EntryTypeName.WIKTIONARY_TRANSLITERATION);
}
//}
} else if (functionName.equals("qualifier")) {
- String qualifier = args.get(0);
- if (!namedArgs.isEmpty() || args.size() > 1) {
- LOG.warning("weird qualifier: " + line);
+ if (args.size() == 0) {
+ otherText.append(wikiTokenizer.token());
+ } else {
+ String qualifier = args.get(0);
+ if (!namedArgs.isEmpty() || args.size() > 1) {
+ LOG.warning("weird qualifier: " + line);
+ }
+ // Unindexed!
+ otherText.append("(").append(qualifier).append(")");
}
- // Unindexed!
- otherText.append("(").append(qualifier).append(")");
} else if (encodings.contains(functionName)) {
otherText.append("").append(args.get(0));
otherIndexBuilder.addEntryWithString(indexedEntry, args.get(0), EntryTypeName.WIKTIONARY_TRANSLATION_OTHER_TEXT);
// American sign language
} else {
// Unindexed!
- otherText.append(wikiTokenizer.token());
+ namedArgs.keySet().removeAll(USELESS_WIKI_ARGS);
+ WikiTokenizer.appendFunction(otherText.append("{{"), functionName, args, namedArgs).append("}}");
}
} else if (wikiTokenizer.isNewline()) {
return;
}
+ if (lang != null) {
+ otherText.insert(0, String.format("(%s) ", lang));
+ }
+
StringBuilder englishText = new StringBuilder();
englishText.append(title);
// -------------------------------------------------------------------------
- private void doForeignWord(final String title, final String text) {
+ private void doForeignWord(final String lang, final String title, final String text) {
final WikiTokenizer wikiTokenizer = new WikiTokenizer(text);
while (wikiTokenizer.nextToken() != null) {
if (wikiTokenizer.isHeading()) {
} else if (headingName.equals("Pronunciation")) {
//doPronunciation(wikiLineReader);
} else if (partOfSpeechHeader.matcher(headingName).matches()) {
- doForeignPartOfSpeech(title, headingName, wikiTokenizer.headingDepth(), wikiTokenizer);
+ doForeignPartOfSpeech(lang, title, headingName, wikiTokenizer.headingDepth(), wikiTokenizer);
}
} else {
}
int foreignCount = 0;
- private void doForeignPartOfSpeech(String title, final String posHeading, final int posDepth, WikiTokenizer wikiTokenizer) {
+ private void doForeignPartOfSpeech(final String lang, String title, final String posHeading, final int posDepth, WikiTokenizer wikiTokenizer) {
if (++foreignCount % 1000 == 0) {
- LOG.info("***" + title + ", pos=" + posHeading + ", foreignCount=" + foreignCount);
+ LOG.info("***" + lang + ", " + title + ", pos=" + posHeading + ", foreignCount=" + foreignCount);
}
if (title.equals("moro")) {
System.out.println();
}
+ boolean titleAppended = false;
final StringBuilder foreignBuilder = new StringBuilder();
final Collection<String> wordForms = new ArrayList<String>();
final List<ListSection> listSections = new ArrayList<ListSection>();
}
if (currentHeadingDepth > posDepth) {
- // TODO
+ // TODO: deal with other neat info sections
continue;
}
// 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!)
// for the conjugation table from "fa".
// Would like to be able to link to a lang#token.
- if (isGender(name)) {
- appendGender(foreignBuilder, name, args);
- } else if (name.equals("wikipedia")) {
- namedArgs.remove("lang");
- if (args.size() > 1 || !namedArgs.isEmpty()) {
- // Unindexed!
- foreignBuilder.append(wikiTokenizer.token());
- } else if (args.size() == 1) {
- foreignBuilder.append(wikiTokenizer.token());
- } else {
- //foreignBuilder.append(title);
- }
- } else if (name.equals("it-noun")) {
+ if (isGender(name)) {
+ appendGender(foreignBuilder, name, args);
+ } else if (name.equals("wikipedia")) {
+ namedArgs.remove("lang");
+ if (args.size() > 1 || !namedArgs.isEmpty()) {
+ // Unindexed!
+ foreignBuilder.append(wikiTokenizer.token());
+ } else if (args.size() == 1) {
+ foreignBuilder.append(wikiTokenizer.token());
+ } else {
+ //foreignBuilder.append(title);
+ }
+ } else if (name.equals("attention") || name.equals("zh-attention")) {
+ // See: http://en.wiktionary.org/wiki/Template:attention
+ // Ignore these.
+ } else if (name.equals("infl")) {
+ // See: http://en.wiktionary.org/wiki/Template:infl
+ final String langCode = get(args, 0);
+ String head = namedArgs.remove("head");
+ if (head == null) {
+ head = namedArgs.remove("title"); // Bug
+ }
+ if (head == null) {
+ head = title;
+ } else {
+ head = WikiTokenizer.toPlainText(head);
+ }
+ titleAppended = true;
+
+ namedArgs.keySet().removeAll(USELESS_WIKI_ARGS);
+
+ final String tr = namedArgs.remove("tr");
+ String g = namedArgs.remove("g");
+ if (g == null) {
+ g = namedArgs.remove("gender");
+ }
+ final String g2 = namedArgs.remove("g2");
+ final String g3 = namedArgs.remove("g3");
+
+ foreignBuilder.append(head);
+
+ if (g != null) {
+ foreignBuilder.append(" {").append(g);
+ if (g2 != null) {
+ foreignBuilder.append("|").append(g2);
+ }
+ if (g3 != null) {
+ foreignBuilder.append("|").append(g3);
+ }
+ foreignBuilder.append("}");
+ }
+
+ if (tr != null) {
+ foreignBuilder.append(String.format(TRANSLITERATION_FORMAT, tr));
+ wordForms.add(tr);
+ }
+
+ final String pos = get(args, 1);
+ if (pos != null) {
+ foreignBuilder.append(" (").append(pos).append(")");
+ }
+ for (int i = 2; i < args.size(); i += 2) {
+ final String inflName = get(args, i);
+ final String inflValue = get(args, i + 1);
+ foreignBuilder.append(", ").append(WikiTokenizer.toPlainText(inflName));
+ if (inflValue != null && inflValue.length() > 0) {
+ foreignBuilder.append(": ").append(WikiTokenizer.toPlainText(inflValue));
+ wordForms.add(inflValue);
+ }
+ }
+ for (final String key : namedArgs.keySet()) {
+ final String value = WikiTokenizer.toPlainText(namedArgs.get(key));
+ foreignBuilder.append(" ").append(key).append("=").append(value);
+ wordForms.add(value);
+ }
+ } else if (name.equals("it-noun")) {
+ titleAppended = true;
final String base = get(args, 0);
final String gender = get(args, 1);
final String singular = base + get(args, 2);
foreignBuilder.append(String.format(" %s {%s}, %s {pl}", singular, gender, plural, plural));
wordForms.add(singular);
wordForms.add(plural);
+ if (!namedArgs.isEmpty() || args.size() > 4) {
+ LOG.warning("Invalid it-noun: " + wikiTokenizer.token());
+ }
} else if (name.equals("it-proper noun")) {
foreignBuilder.append(wikiTokenizer.token());
} else if (name.equals("it-adj")) {
foreignBuilder.append(wikiTokenizer.token());
// LOG.warning("Unknown function: " + wikiTokenizer.token());
}
-
} else if (wikiTokenizer.isListItem()) {
final String prefix = wikiTokenizer.listItemPrefix();
if (lastListSection != null &&
} finally {
// Here's where we exit.
- // TODO: Should we make an entry even if there are no foreign list items?
+ // Should we make an entry even if there are no foreign list items?
String foreign = foreignBuilder.toString().trim();
- if (!foreign.toLowerCase().startsWith(title.toLowerCase())) {
- foreign = title + " " + foreign;
+ if (!titleAppended && !foreign.toLowerCase().startsWith(title.toLowerCase())) {
+ foreign = String.format("%s %s", title, foreign);
+ }
+ if (!langPattern.matcher(lang).matches()) {
+ foreign = String.format("(%s) %s", lang, foreign);
}
for (final ListSection listSection : listSections) {
doForeignListItem(foreign, title, wordForms, listSection);
"imperative"
);
+ // Might only want to remove "lang" if it's equal to "zh", for example.
+ static final Set<String> USELESS_WIKI_ARGS = new LinkedHashSet<String>(Arrays.asList("lang", "sc", "sort", "cat"));
private void doForeignListItem(final String foreignText, String title, final Collection<String> forms, final ListSection listSection) {
final String prefix = listSection.firstPrefix;
if (prefix.length() > 1) {
+ // Could just get looser and say that any prefix longer than first is a sublist.
LOG.warning("Prefix too long: " + listSection);
return;
}
final StringBuilder englishBuilder = new StringBuilder();
final String mainLine = listSection.firstLine;
-
final WikiTokenizer englishTokenizer = new WikiTokenizer(mainLine, false);
while (englishTokenizer.nextToken() != null) {
// TODO handle form of....
}
} else if (englishTokenizer.isFunction()) {
final String name = englishTokenizer.functionName();
- if (name.contains("conjugation of ") ||
- name.contains("form of ") ||
- name.contains("feminine of ") ||
- name.contains("plural of ")) {
- // Ignore these in the index, they're really annoying....
- englishBuilder.append(englishTokenizer.token());
+ final List<String> args = englishTokenizer.functionPositionArgs();
+ final Map<String,String> namedArgs = englishTokenizer.functionNamedArgs();
+
+ if (
+ name.equals("form of") ||
+ name.contains("conjugation of") ||
+ name.contains("participle of") ||
+ name.contains("gerund of") ||
+ name.contains("feminine of") ||
+ name.contains("plural of")) {
+ String formName = name;
+ if (name.equals("form of")) {
+ formName = args.remove(0);
+ }
+ String baseForm = get(args, 1, "");
+ if ("".equals(baseForm)) {
+ baseForm = get(args, 0, null);
+ remove(args, 1, "");
+ } else {
+ remove(args, 0, null);
+ }
+ namedArgs.keySet().removeAll(USELESS_WIKI_ARGS);
+ WikiTokenizer.appendFunction(englishBuilder.append("{"), formName, args, namedArgs).append("}");
+ otherIndexBuilder.addEntryWithString(indexedEntry, baseForm, EntryTypeName.WIKTIONARY_BASE_FORM_SINGLE, EntryTypeName.WIKTIONARY_BASE_FORM_MULTI);
+
} else {
- englishBuilder.append(englishTokenizer.token());
+ namedArgs.keySet().removeAll(USELESS_WIKI_ARGS);
+ if (args.size() == 0 && namedArgs.isEmpty()) {
+ englishBuilder.append("{").append(name).append("}");
+ } else {
+ WikiTokenizer.appendFunction(englishBuilder.append("{{"), name, args, namedArgs).append("}}");
+ }
// LOG.warning("Unexpected function: " + englishTokenizer.token());
}
} else {
if (englishTokenizer.isComment() || englishTokenizer.isMarkup()) {
} else {
- LOG.warning("Unexpected definition text: " + englishTokenizer.token());
+ LOG.warning("Unexpected definition type: " + englishTokenizer.token());
}
}
}
pairEntry.pairs.add(pair);
otherIndexBuilder.addEntryWithString(indexedEntry, title, EntryTypeName.WIKTIONARY_TITLE_SINGLE, EntryTypeName.WIKTIONARY_TITLE_MULTI);
for (final String form : forms) {
- otherIndexBuilder.addEntryWithString(indexedEntry, form, EntryTypeName.WIKTIONARY_FORM_SINGLE, EntryTypeName.WIKTIONARY_FORM_MULTI);
+ otherIndexBuilder.addEntryWithString(indexedEntry, form, EntryTypeName.WIKTIONARY_INFLECTD_FORM_SINGLE, EntryTypeName.WIKTIONARY_INFLECTED_FORM_MULTI);
}
}
mdashLen = 3;
}
- // TODO: index and clean these!!!
- if (nextPrefix.equals("#:") && dash != -1) {
+ if ((nextPrefix.equals("#:") || nextPrefix.equals("##:")) && dash != -1) {
final String foreignEx = nextLine.substring(0, dash);
final String englishEx = nextLine.substring(dash + mdashLen);
- final Pair pair = new Pair(formatAndIndexExampleString(englishEx, enIndexBuilder), formatAndIndexExampleString(foreignEx, otherIndexBuilder), swap);
- pairEntry.pairs.add(pair);
+ final Pair pair = new Pair(formatAndIndexExampleString(englishEx, enIndexBuilder, indexedEntry), formatAndIndexExampleString(foreignEx, otherIndexBuilder, indexedEntry), swap);
+ if (pair.lang1 != "--" && pair.lang1 != "--") {
+ pairEntry.pairs.add(pair);
+ }
lastForeign = null;
- } else if (nextPrefix.equals("#:")){
- final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null), swap);
+ } else if (nextPrefix.equals("#:") || nextPrefix.equals("##:")){
+ final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
lastForeign = nextLine;
- pairEntry.pairs.add(pair);
+ if (pair.lang1 != "--" && pair.lang1 != "--") {
+ pairEntry.pairs.add(pair);
+ }
} else if (nextPrefix.equals("#::") || nextPrefix.equals("#**")) {
- if (lastForeign != null) {
+ if (lastForeign != null && pairEntry.pairs.size() > 0) {
pairEntry.pairs.remove(pairEntry.pairs.size() - 1);
- final Pair pair = new Pair(formatAndIndexExampleString(nextLine, enIndexBuilder), formatAndIndexExampleString(lastForeign, otherIndexBuilder), swap);
- pairEntry.pairs.add(pair);
+ final Pair pair = new Pair(formatAndIndexExampleString(nextLine, enIndexBuilder, indexedEntry), formatAndIndexExampleString(lastForeign, otherIndexBuilder, indexedEntry), swap);
+ if (pair.lang1 != "--" || pair.lang2 != "--") {
+ pairEntry.pairs.add(pair);
+ }
+ lastForeign = null;
} else {
- LOG.warning("English example with no foreign: " + title + ", " + nextLine);
+ LOG.warning("TODO: English example with no foreign: " + title + ", " + nextLine);
+ final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
+ if (pair.lang1 != "--" || pair.lang2 != "--") {
+ pairEntry.pairs.add(pair);
+ }
}
} else if (nextPrefix.equals("#*")) {
// Can't really index these.
- final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null), swap);
+ final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
lastForeign = nextLine;
- pairEntry.pairs.add(pair);
- } else if (nextPrefix.equals("#::*")) {
- final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null), swap);
- pairEntry.pairs.add(pair);
- } else {
- assert false;
+ if (pair.lang1 != "--" || pair.lang2 != "--") {
+ pairEntry.pairs.add(pair);
+ }
+ } else if (nextPrefix.equals("#::*") || nextPrefix.equals("##") || nextPrefix.equals("#*:") || nextPrefix.equals("#:*") || true) {
+ final Pair pair = new Pair("--", formatAndIndexExampleString(nextLine, null, indexedEntry), swap);
+ if (pair.lang1 != "--" || pair.lang2 != "--") {
+ pairEntry.pairs.add(pair);
+ }
+// } else {
+// assert false;
}
}
}
- private String formatAndIndexExampleString(final String example, final IndexBuilder indexBuilder) {
+ private String formatAndIndexExampleString(final String example, final IndexBuilder indexBuilder, final IndexedEntry indexedEntry) {
final WikiTokenizer wikiTokenizer = new WikiTokenizer(example, false);
final StringBuilder builder = new StringBuilder();
boolean insideTripleQuotes = false;
while (wikiTokenizer.nextToken() != null) {
if (wikiTokenizer.isPlainText()) {
builder.append(wikiTokenizer.token());
-
+ if (indexBuilder != null) {
+ indexBuilder.addEntryWithString(indexedEntry, wikiTokenizer.token(), EntryTypeName.WIKTIONARY_EXAMPLE);
+ }
} else if (wikiTokenizer.isWikiLink()) {
- builder.append(wikiTokenizer.wikiLinkText());
-
+ final String text = wikiTokenizer.wikiLinkText().replaceAll("'", "");
+ builder.append(text);
+ if (indexBuilder != null) {
+ indexBuilder.addEntryWithString(indexedEntry, text, EntryTypeName.WIKTIONARY_EXAMPLE);
+ }
} else if (wikiTokenizer.isFunction()) {
builder.append(wikiTokenizer.token());
} else if (wikiTokenizer.isMarkup()) {
LOG.warning("unexpected token: " + wikiTokenizer.token());
}
}
- return trim(builder.toString());
+ final String result = trim(builder.toString());
+ return result.length() > 0 ? result : "--";
}