import java.util.List;
import java.util.Map;
import java.util.Set;
+import java.util.logging.Level;
import java.util.logging.Logger;
import java.util.regex.Pattern;
+import com.hughes.android.dictionary.engine.EntrySource;
import com.hughes.android.dictionary.engine.EntryTypeName;
import com.hughes.android.dictionary.engine.IndexBuilder;
import com.hughes.android.dictionary.engine.IndexedEntry;
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 {
- private static final String TRANSLITERATION_FORMAT = " (tr. %s)";
-
static final Logger LOG = Logger.getLogger(EnWiktionaryXmlParser.class.getName());
// TODO: process {{ttbc}} lines
"Preposition|Proper noun|Article|Prepositional phrase|Acronym|" +
"Abbreviation|Initialism|Contraction|Prefix|Suffix|Symbol|Letter|" +
"Ligature|Idiom|Phrase|\\{\\{acronym\\}\\}|\\{\\{initialism\\}\\}|" +
+ "\\{\\{abbreviation\\}\\}|" +
// These are @deprecated:
"Noun form|Verb form|Adjective form|Nominal phrase|Noun phrase|" +
"Verb phrase|Transitive verb|Intransitive verb|Reflexive verb|" +
"Particle|Interjection|Pronominal adverb" +
"Han character|Hanzi|Hanja|Kanji|Katakana character|Syllable");
+ EntrySource entrySource;
final IndexBuilder enIndexBuilder;
final IndexBuilder foreignIndexBuilder;
final Pattern langPattern;
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) {
}
- public void parse(final File file, final int pageLimit) throws IOException {
+ public void parse(final File file, final EntrySource entrySource, final int pageLimit) throws IOException {
+ this.entrySource = entrySource;
int pageCount = 0;
final DataInputStream dis = new DataInputStream(new BufferedInputStream(new FileInputStream(file)));
while (true) {
try {
title = dis.readUTF();
} catch (EOFException e) {
- LOG.warning("Error reading split!");
+ LOG.log(Level.INFO, "EOF reading split.");
dis.close();
return;
}
} else if (headerName.equals("Translations")) {
if (pos == null) {
- LOG.warning("Translations without POS: " + title);
+ LOG.info("Translations without POS (but using anyway): " + title);
}
doTranslations(wikiTokenizer, pos);
} else if (headerName.equals("Pronunciation")) {
}
} else if (wikiTokenizer.isFunction()) {
final String name = wikiTokenizer.functionName();
- if (name.equals("head")) {
+ if (name.equals("head") && pos == null) {
LOG.warning("{{head}} without POS: " + title);
}
}
}
final AppendAndIndexWikiCallback appendAndIndexWikiCallback = new AppendAndIndexWikiCallback(this);
+ {
+ appendAndIndexWikiCallback.functionCallbacks.putAll(FunctionCallbacksDefault.DEFAULT);
+ }
private void doTranslations(final WikiTokenizer wikiTokenizer, final String pos) {
if (title.equals("absolutely")) {
}
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 PairEntry pairEntry = new PairEntry(entrySource);
final IndexedEntry indexedEntry = new IndexedEntry(pairEntry);
final StringBuilder foreignText = new StringBuilder();
appendAndIndexWikiCallback.reset(foreignText, indexedEntry);
- appendAndIndexWikiCallback.functionCallbacks.clear();
- appendAndIndexWikiCallback.functionCallbacks.putAll(FunctionCallbacksDefault.DEFAULT);
appendAndIndexWikiCallback.dispatch(rest, foreignIndexBuilder, EntryTypeName.WIKTIONARY_TRANSLATION_OTHER_TEXT);
if (foreignText.length() == 0) {
if (!pairsAdded.add(pair.toString())) {
LOG.warning("Duplicate pair: " + pair.toString());
}
- if (pair.toString().equals("libero {m} :: free (adjective)")) {
- System.out.println();
- }
-
}
int foreignCount = 0;
+ 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 (FunctionCallbacksDefault.DEFAULT.get(name) instanceof FunctionCallbacksDefault.Gender) {
- // TODO: Fix hack!
- appendAndIndexWikiCallback.reset(foreignBuilder, null);
- FunctionCallbacksDefault.DEFAULT.get(name).onWikiFunction(wikiTokenizer, name, args, namedArgs, this, appendAndIndexWikiCallback);
- } 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 = ListUtil.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 = ListUtil.get(args, 1);
- if (pos != null) {
- foreignBuilder.append(" (").append(pos).append(")");
- }
- for (int i = 2; i < args.size(); i += 2) {
- final String inflName = ListUtil.get(args, i);
- final String inflValue = ListUtil.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 = ListUtil.get(args, 0);
- final String gender = ListUtil.get(args, 1);
- final String singular = base + ListUtil.get(args, 2, null);
- final String plural = base + ListUtil.get(args, 3, null);
- 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 &&
foreign = String.format("(%s) %s", lang, foreign);
}
for (final ListSection listSection : listSections) {
- doForeignListItem(foreign, title, wordForms, listSection);
+ doForeignListSection(foreign, title, wordForms, listSection);
}
}
}
- 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(
"sc",
"sort",
"cat",
- "xs"));
+ "xs",
+ "nodot"));
- private void doForeignListItem(final String foreignText, String title, final Collection<String> forms, final ListSection listSection) {
-
+ public boolean entryIsFormOfSomething = false;
+
+ private void doForeignListSection(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.
return;
}
- final PairEntry pairEntry = new PairEntry();
+ final PairEntry pairEntry = new PairEntry(entrySource);
final IndexedEntry indexedEntry = new IndexedEntry(pairEntry);
-
- final StringBuilder englishBuilder = new StringBuilder();
+ entryIsFormOfSomething = false;
+ 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.addEntryWithStringNoSingle(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.addEntryWithStringNoSingle(indexedEntry, text, EntryTypeName.WIKTIONARY_ENGLISH_DEF_WIKI_LINK);
- } else if (link.contains("#") && this.langPattern.matcher(link).find()) {
- englishBuilder.append(text);
- foreignIndexBuilder.addEntryWithStringNoSingle(indexedEntry, text, EntryTypeName.WIKTIONARY_ENGLISH_DEF_OTHER_LANG);
- } else if (link.equals("plural")) {
- englishBuilder.append(text);
- } else {
- //LOG.warning("Special link: " + englishTokenizer.token());
- enIndexBuilder.addEntryWithStringNoSingle(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.addEntryWithStringNoSingle(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 = ListUtil.remove(args, 0, null);
- }
- if (formName == null) {
- LOG.warning("Missing form name: " + title);
- formName = "form of";
- }
- String baseForm = ListUtil.get(args, 1, "");
- if ("".equals(baseForm)) {
- baseForm = ListUtil.get(args, 0, null);
- ListUtil.remove(args, 1, "");
- } else {
- ListUtil.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_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.addEntryWithStringNoSingle(indexedEntry, args.get(1), EntryTypeName.WIKTIONARY_ENGLISH_DEF_WIKI_LINK);
- } else {
- foreignIndexBuilder.addEntryWithStringNoSingle(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_MULTI);
+ foreignIndexBuilder.addEntryWithString(indexedEntry, title, entryIsFormOfSomething ? EntryTypeName.WIKTIONARY_IS_FORM_OF_SOMETHING_ELSE : EntryTypeName.WIKTIONARY_TITLE_MULTI);
for (final String form : forms) {
foreignIndexBuilder.addEntryWithString(indexedEntry, form, EntryTypeName.WIKTIONARY_INFLECTED_FORM_MULTI);
}
for (int i = 0; i < listSection.nextPrefixes.size(); ++i) {
final String nextPrefix = listSection.nextPrefixes.get(i);
final String nextLine = listSection.nextLines.get(i);
+
+ // TODO: This splitting is not sensitive to wiki code.
int dash = nextLine.indexOf("—");
int mdashLen = 7;
if (dash == -1) {
}
private String formatAndIndexExampleString(final String example, final IndexBuilder indexBuilder, final IndexedEntry indexedEntry) {
- final WikiTokenizer wikiTokenizer = new WikiTokenizer(example, false);
+ // TODO:
+// if (wikiTokenizer.token().equals("'''")) {
+// insideTripleQuotes = !insideTripleQuotes;
+// }
final StringBuilder builder = new StringBuilder();
- boolean insideTripleQuotes = false;
- while (wikiTokenizer.nextToken() != null) {
- if (wikiTokenizer.isPlainText()) {
- builder.append(wikiTokenizer.token());
- if (indexBuilder != null) {
- 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.addEntryWithStringNoSingle(indexedEntry, text, EntryTypeName.WIKTIONARY_EXAMPLE);
- }
- } else if (wikiTokenizer.isFunction()) {
- builder.append(wikiTokenizer.token());
- } else if (wikiTokenizer.isMarkup()) {
- if (wikiTokenizer.token().equals("'''")) {
- insideTripleQuotes = !insideTripleQuotes;
- }
- } else if (wikiTokenizer.isComment() || wikiTokenizer.isNewline()) {
- // Do nothing.
- } else {
- LOG.warning("unexpected token: " + wikiTokenizer.token());
- }
+ appendAndIndexWikiCallback.reset(builder, indexedEntry);
+ appendAndIndexWikiCallback.entryTypeName = EntryTypeName.WIKTIONARY_EXAMPLE;
+ appendAndIndexWikiCallback.entryTypeNameSticks = true;
+ try {
+ // TODO: this is a hack needed because we don't safely split on the dash.
+ appendAndIndexWikiCallback.dispatch(example, indexBuilder, EntryTypeName.WIKTIONARY_EXAMPLE);
+ } catch (AssertionError e) {
+ return "--";
}
final String result = trim(builder.toString());
return result.length() > 0 ? result : "--";
projects / DictionaryPC.git / blobdiff