import com.hughes.android.dictionary.engine.PairEntry;
import com.hughes.android.dictionary.engine.PairEntry.Pair;
import com.hughes.android.dictionary.parser.WikiTokenizer;
+import com.hughes.util.ListUtil;
public class EnWiktionaryXmlParser {
final Pattern langPattern;
final Pattern langCodePattern;
final boolean swap;
+
+ // State used while parsing.
+ enum State {
+ TRANSLATION_LINE,
+ ENGLISH_DEF_OF_FOREIGN,
+ ENGLISH_EXAMPLE,
+ FOREIGN_EXAMPLE,
+ }
+ State state = null;
+ String title;
public EnWiktionaryXmlParser(final IndexBuilder enIndexBuilder, final IndexBuilder otherIndexBuilder, final Pattern langPattern, final Pattern langCodePattern, final boolean swap) {
this.enIndexBuilder = enIndexBuilder;
return;
}
- final String title;
try {
title = dis.readUTF();
} catch (EOFException e) {
+ LOG.warning("Error reading split!");
dis.close();
return;
}
dis.readFully(bytes);
final String text = new String(bytes, "UTF8");
- parseSection(title, heading, text);
+ parseSection(heading, text);
++pageCount;
if (pageCount % 1000 == 0) {
}
}
- private void parseSection(final String title, String heading, final String text) {
+ private void parseSection(String heading, final String text) {
if (title.startsWith("Wiktionary:") ||
title.startsWith("Template:") ||
title.startsWith("Appendix:") ||
heading = heading.replaceAll("=", "").trim();
if (heading.equals("English")) {
- doEnglishWord(title, text);
+ doEnglishWord(text);
} else if (langPattern.matcher(heading).find()){
- doForeignWord(heading, title, text);
+ doForeignWord(heading, text);
}
} // endPage()
// -------------------------------------------------------------------------
- private void doEnglishWord(String title, String text) {
+ private void doEnglishWord(String text) {
String pos = null;
int posDepth = -1;
if (pos == null) {
LOG.warning("Translations without POS: " + title);
}
- doTranslations(title, wikiTokenizer, pos);
+ doTranslations(wikiTokenizer, pos);
} else if (headerName.equals("Pronunciation")) {
//doPronunciation(wikiLineReader);
}
}
}
}
-
-
- private static Set<String> encodings = new LinkedHashSet<String>(Arrays.asList("zh-ts",
- "sd-Arab", "ku-Arab", "Arab", "unicode", "Laoo", "ur-Arab", "Thai",
- "fa-Arab", "Khmr", "zh-tsp", "Cyrl", "IPAchar", "ug-Arab", "ko-inline",
- "Jpan", "Kore", "Hebr", "rfscript", "Beng", "Mong", "Knda", "Cyrs",
- "yue-tsj", "Mlym", "Tfng", "Grek", "yue-yue-j"));
- private void doTranslations(final String title, final WikiTokenizer wikiTokenizer, final String pos) {
+ final AppendAndIndexWikiCallback appendAndIndexWikiCallback = new AppendAndIndexWikiCallback(this);
+ {
+ appendAndIndexWikiCallback.functionCallbacks.putAll(FunctionCallbacksDefault.DEFAULT);
+ }
+
+ private void doTranslations(final WikiTokenizer wikiTokenizer, final String pos) {
if (title.equals("absolutely")) {
//System.out.println();
}
String rest = line.substring(colonIndex + 1).trim();
if (rest.length() > 0) {
- doTranslationLine(line, appendLang ? lang : null, title, pos, sense, rest);
+ doTranslationLine(line, appendLang ? lang : null, pos, sense, rest);
}
} else if (wikiTokenizer.remainderStartsWith("''See''")) {
}
}
- private void doTranslationLine(final String line, final String lang, final String title, final String pos, final String sense, final String rest) {
+ private void doTranslationLine(final String line, final String lang, final String pos, final String sense, final String rest) {
+ state = State.TRANSLATION_LINE;
// Good chance we'll actually file this one...
final PairEntry pairEntry = new PairEntry();
final IndexedEntry indexedEntry = new IndexedEntry(pairEntry);
final StringBuilder foreignText = new StringBuilder();
- final WikiTokenizer wikiTokenizer = new WikiTokenizer(rest, false);
- while (wikiTokenizer.nextToken() != null) {
-
- if (wikiTokenizer.isPlainText()) {
- final String plainText = wikiTokenizer.token();
- foreignText.append(plainText);
- foreignIndexBuilder.addEntryWithString(indexedEntry, plainText, EntryTypeName.WIKTIONARY_TRANSLATION_OTHER_TEXT);
-
- } else if (wikiTokenizer.isWikiLink()) {
- final String plainText = wikiTokenizer.wikiLinkText();
- foreignText.append(plainText);
- // TODO: should check for English before appending.
- foreignIndexBuilder.addEntryWithString(indexedEntry, plainText, EntryTypeName.WIKTIONARY_TRANSLATION_WIKI_TEXT);
-
- } else if (wikiTokenizer.isFunction()) {
- final String functionName = wikiTokenizer.functionName();
- final List<String> args = wikiTokenizer.functionPositionArgs();
- final Map<String,String> namedArgs = wikiTokenizer.functionNamedArgs();
-
- if (functionName.equals("t") || functionName.equals("t+") || functionName.equals("t-") || functionName.equals("tø") || functionName.equals("apdx-t")) {
- } else if (functionName.equals("qualifier")) {
- } else if (encodings.contains(functionName)) {
- foreignText.append("").append(args.get(0));
- foreignIndexBuilder.addEntryWithString(indexedEntry, args.get(0), EntryTypeName.WIKTIONARY_TRANSLATION_OTHER_TEXT);
- } else if (isGender(functionName)) {
- appendGender(foreignText, functionName, args);
- } else if (functionName.equals("g")) {
- foreignText.append("{g}");
- } else if (functionName.equals("l")) {
- // encodes text in various langs.
- // lang is arg 0.
- foreignText.append("").append(args.get(1));
- foreignIndexBuilder.addEntryWithString(indexedEntry, args.get(1), EntryTypeName.WIKTIONARY_TRANSLATION_OTHER_TEXT);
- // TODO: transliteration
- } else if (functionName.equals("term")) {
- // cross-reference to another dictionary
- foreignText.append("").append(args.get(0));
- foreignIndexBuilder.addEntryWithString(indexedEntry, args.get(0), EntryTypeName.WIKTIONARY_TRANSLATION_OTHER_TEXT);
- // TODO: transliteration
- } else if (functionName.equals("italbrac") || functionName.equals("gloss")) {
- // TODO: put this text aside to use it.
- foreignText.append("[").append(args.get(0)).append("]");
- foreignIndexBuilder.addEntryWithString(indexedEntry, args.get(0), EntryTypeName.WIKTIONARY_TRANSLATION_OTHER_TEXT);
- } else if (functionName.equals("ttbc")) {
- LOG.warning("Unexpected {{ttbc}}");
- } else if (functionName.equals("trreq")) {
- } else if (functionName.equals("not used")) {
- foreignText.append("(not used)");
- } else if (functionName.equals("t-image")) {
- // American sign language
- } else {
- // Unindexed!
- namedArgs.keySet().removeAll(USELESS_WIKI_ARGS);
- WikiTokenizer.appendFunction(foreignText.append("{{"), functionName, args, namedArgs).append("}}");
- }
-
- } else if (wikiTokenizer.isNewline()) {
- } else if (wikiTokenizer.isComment()) {
- } else if (wikiTokenizer.isMarkup()) {
- } else {
- LOG.warning("Bad translation token: " + wikiTokenizer.token());
- }
- } // while-token loop.
-
+ appendAndIndexWikiCallback.reset(foreignText, indexedEntry);
+ appendAndIndexWikiCallback.dispatch(rest, foreignIndexBuilder, EntryTypeName.WIKTIONARY_TRANSLATION_OTHER_TEXT);
if (foreignText.length() == 0) {
LOG.warning("Empty foreignText: " + line);
englishText.append(title);
if (sense != null) {
englishText.append(" (").append(sense).append(")");
- enIndexBuilder.addEntryWithString(indexedEntry, sense, EntryTypeName.WIKTIONARY_TRANSLATION_SENSE, EntryTypeName.WIKTIONARY_TRANSLATION_SENSE);
+ enIndexBuilder.addEntryWithString(indexedEntry, sense, EntryTypeName.WIKTIONARY_TRANSLATION_SENSE);
}
if (pos != null) {
englishText.append(" (").append(pos.toLowerCase()).append(")");
}
- enIndexBuilder.addEntryWithString(indexedEntry, title, EntryTypeName.WIKTIONARY_TITLE_SINGLE, EntryTypeName.WIKTIONARY_TITLE_MULTI);
+ enIndexBuilder.addEntryWithString(indexedEntry, title, EntryTypeName.WIKTIONARY_TITLE_MULTI);
final Pair pair = new Pair(trim(englishText.toString()), trim(foreignText.toString()), swap);
pairEntry.pairs.add(pair);
}
- private void appendGender(final StringBuilder otherText,
- final String functionName, final List<String> args) {
- otherText.append("{");
- otherText.append(functionName);
- for (int i = 0; i < args.size(); ++i) {
- otherText.append("|").append(args.get(i));
- }
- otherText.append("}");
- }
-
-
- private boolean isGender(final String functionName) {
- return functionName.equals("m") || functionName.equals("f") || functionName.equals("n") || functionName.equals("p");
- }
-
Set<String> pairsAdded = new LinkedHashSet<String>();
// -------------------------------------------------------------------------
- private void doForeignWord(final String lang, final String title, final String text) {
+ private void doForeignWord(final String lang, 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(lang, title, headingName, wikiTokenizer.headingDepth(), wikiTokenizer);
+ doForeignPartOfSpeech(lang, headingName, wikiTokenizer.headingDepth(), wikiTokenizer);
}
} else {
}
int foreignCount = 0;
- private void doForeignPartOfSpeech(final String lang, String title, final String posHeading, final int posDepth, WikiTokenizer wikiTokenizer) {
+ final Collection<String> wordForms = new ArrayList<String>();
+ boolean titleAppended = false;
+
+
+ private void doForeignPartOfSpeech(final String lang, String posHeading, final int posDepth, WikiTokenizer wikiTokenizer) {
if (++foreignCount % 1000 == 0) {
LOG.info("***" + lang + ", " + title + ", pos=" + posHeading + ", foreignCount=" + foreignCount);
}
- if (title.equals("moro")) {
+ if (title.equals("6")) {
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>();
+ appendAndIndexWikiCallback.reset(foreignBuilder, null);
+ this.state = State.ENGLISH_DEF_OF_FOREIGN; // TODO: this is wrong, need new category....
+ titleAppended = false;
+ wordForms.clear();
+
try {
ListSection lastListSection = null;
// 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("attention") || name.equals("zh-attention")) {
- // See: http://en.wiktionary.org/wiki/Template:attention
- // Ignore these.
- } else if (name.equals("infl") || name.equals("head")) {
- // 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);
- final String plural = base + get(args, 3);
- 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());
- } else if (name.startsWith("it-conj")) {
- if (name.equals("it-conj-are")) {
- itConjAre(args, namedArgs);
- } else if (name.equals("it-conj-ere")) {
- } else if (name.equals("it-conj-ire")) {
- } else {
- LOG.warning("Unknown conjugation: " + wikiTokenizer.token());
- }
- } else {
- // Unindexed!
- foreignBuilder.append(wikiTokenizer.token());
- // LOG.warning("Unknown function: " + wikiTokenizer.token());
- }
+
+ appendAndIndexWikiCallback.onFunction(wikiTokenizer, name, args, namedArgs);
+
} else if (wikiTokenizer.isListItem()) {
final String prefix = wikiTokenizer.listItemPrefix();
if (lastListSection != null &&
}
- static final Pattern UNINDEXED_WIKI_TEXT = Pattern.compile(
- "(first|second|third)-person (singular|plural)|" +
- "present tense|" +
- "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"));
+ static final Set<String> USELESS_WIKI_ARGS = new LinkedHashSet<String>(
+ Arrays.asList(
+ "lang",
+ "sc",
+ "sort",
+ "cat",
+ "xs"));
private void doForeignListItem(final String foreignText, String title, final Collection<String> forms, final ListSection listSection) {
-
+ state = State.ENGLISH_DEF_OF_FOREIGN;
final String prefix = listSection.firstPrefix;
if (prefix.length() > 1) {
// Could just get looser and say that any prefix longer than first is a sublist.
final IndexedEntry indexedEntry = new IndexedEntry(pairEntry);
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....
- if (englishTokenizer.isPlainText()) {
- englishBuilder.append(englishTokenizer.token());
- enIndexBuilder.addEntryWithString(indexedEntry, englishTokenizer.token(), EntryTypeName.WIKTIONARY_ENGLISH_DEF);
- } else if (englishTokenizer.isWikiLink()) {
- final String text = englishTokenizer.wikiLinkText();
- final String link = englishTokenizer.wikiLinkDest();
- if (link != null) {
- if (link.contains("#English")) {
- englishBuilder.append(text);
- enIndexBuilder.addEntryWithString(indexedEntry, text, EntryTypeName.WIKTIONARY_ENGLISH_DEF_WIKI_LINK);
- } else if (link.contains("#") && this.langPattern.matcher(link).find()) {
- englishBuilder.append(text);
- foreignIndexBuilder.addEntryWithString(indexedEntry, text, EntryTypeName.WIKTIONARY_ENGLISH_DEF_OTHER_LANG);
- } else if (link.equals("plural")) {
- englishBuilder.append(text);
- } else {
- //LOG.warning("Special link: " + englishTokenizer.token());
- enIndexBuilder.addEntryWithString(indexedEntry, text, EntryTypeName.WIKTIONARY_ENGLISH_DEF_WIKI_LINK);
- englishBuilder.append(text);
- }
- } else {
- // link == null
- englishBuilder.append(text);
- if (!UNINDEXED_WIKI_TEXT.matcher(text).find()) {
- enIndexBuilder.addEntryWithString(indexedEntry, text, EntryTypeName.WIKTIONARY_ENGLISH_DEF_WIKI_LINK);
- }
- }
- } else if (englishTokenizer.isFunction()) {
- final String name = englishTokenizer.functionName();
- 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 = remove(args, 0, null);
- }
- if (formName == null) {
- LOG.warning("Missing form name: " + title);
- formName = "form of";
- }
- 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("}");
- if (baseForm != null) {
- foreignIndexBuilder.addEntryWithString(indexedEntry, baseForm, EntryTypeName.WIKTIONARY_BASE_FORM_SINGLE, EntryTypeName.WIKTIONARY_BASE_FORM_MULTI);
- } else {
- // null baseForm happens in Danish.
- LOG.warning("Null baseform: " + title);
- }
- } else if (name.equals("l")) {
- // encodes text in various langs.
- // lang is arg 0.
- englishBuilder.append("").append(args.get(1));
- final String langCode = args.get(0);
- if ("en".equals(langCode)) {
- enIndexBuilder.addEntryWithString(indexedEntry, args.get(1), EntryTypeName.WIKTIONARY_ENGLISH_DEF_WIKI_LINK);
- } else {
- foreignIndexBuilder.addEntryWithString(indexedEntry, args.get(1), EntryTypeName.WIKTIONARY_ENGLISH_DEF_OTHER_LANG);
- }
- // TODO: transliteration
-
- } else if (name.equals("defn") || name.equals("rfdef")) {
- // Do nothing.
- // http://en.wiktionary.org/wiki/Wiktionary:Requests_for_deletion/Others#Template:defn
- // Redundant, used for the same purpose as {{rfdef}}, but this
- // doesn't produce the "This word needs a definition" text.
- // Delete or redirect.
- } else {
- 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 type: " + englishTokenizer.token());
- }
- }
- }
-
+
+ appendAndIndexWikiCallback.reset(englishBuilder, indexedEntry);
+ appendAndIndexWikiCallback.dispatch(mainLine, enIndexBuilder, EntryTypeName.WIKTIONARY_ENGLISH_DEF);
+
final String english = trim(englishBuilder.toString());
if (english.length() > 0) {
final Pair pair = new Pair(english, trim(foreignText), this.swap);
pairEntry.pairs.add(pair);
- foreignIndexBuilder.addEntryWithString(indexedEntry, title, EntryTypeName.WIKTIONARY_TITLE_SINGLE, EntryTypeName.WIKTIONARY_TITLE_MULTI);
+ foreignIndexBuilder.addEntryWithString(indexedEntry, title, EntryTypeName.WIKTIONARY_TITLE_MULTI);
for (final String form : forms) {
- foreignIndexBuilder.addEntryWithString(indexedEntry, form, EntryTypeName.WIKTIONARY_INFLECTD_FORM_SINGLE, EntryTypeName.WIKTIONARY_INFLECTED_FORM_MULTI);
+ foreignIndexBuilder.addEntryWithString(indexedEntry, form, EntryTypeName.WIKTIONARY_INFLECTED_FORM_MULTI);
}
}
if (wikiTokenizer.isPlainText()) {
builder.append(wikiTokenizer.token());
if (indexBuilder != null) {
- indexBuilder.addEntryWithString(indexedEntry, wikiTokenizer.token(), EntryTypeName.WIKTIONARY_EXAMPLE);
+ indexBuilder.addEntryWithStringNoSingle(indexedEntry, wikiTokenizer.token(), EntryTypeName.WIKTIONARY_EXAMPLE);
}
} else if (wikiTokenizer.isWikiLink()) {
final String text = wikiTokenizer.wikiLinkText().replaceAll("'", "");
builder.append(text);
if (indexBuilder != null) {
- indexBuilder.addEntryWithString(indexedEntry, text, EntryTypeName.WIKTIONARY_EXAMPLE);
+ indexBuilder.addEntryWithStringNoSingle(indexedEntry, text, EntryTypeName.WIKTIONARY_EXAMPLE);
}
} else if (wikiTokenizer.isFunction()) {
builder.append(wikiTokenizer.token());