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.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.LinkedHashMap;
24 import java.util.LinkedHashSet;
25 import java.util.List;
28 import java.util.SortedMap;
29 import java.util.TreeMap;
30 import java.util.concurrent.atomic.AtomicInteger;
31 import java.util.logging.Level;
32 import java.util.logging.Logger;
33 import java.util.regex.Pattern;
35 import com.hughes.android.dictionary.engine.EntrySource;
36 import com.hughes.android.dictionary.engine.EntryTypeName;
37 import com.hughes.android.dictionary.engine.IndexBuilder;
38 import com.hughes.android.dictionary.engine.IndexedEntry;
39 import com.hughes.android.dictionary.parser.Parser;
40 import com.hughes.android.dictionary.parser.WikiTokenizer;
41 import com.hughes.util.EnumUtil;
43 public abstract class AbstractWiktionaryParser implements Parser {
45 static final Logger LOG = Logger.getLogger("WiktionaryParser");
47 final SortedMap<String, AtomicInteger> counters = new TreeMap<String, AtomicInteger>();
48 final Set<String> pairsAdded = new LinkedHashSet<String>();
50 EntrySource entrySource;
54 abstract void parseSection(final String heading, final String text);
57 public void parse(final File file, final EntrySource entrySource, final int pageLimit) throws IOException {
58 this.entrySource = entrySource;
60 final DataInputStream dis = new DataInputStream(new BufferedInputStream(new FileInputStream(file)));
63 if (pageLimit >= 0 && pageCount >= pageLimit) {
68 title = dis.readUTF();
69 } catch (EOFException e) {
70 LOG.log(Level.INFO, "EOF reading split.");
74 final String heading = dis.readUTF();
75 final int bytesLength = dis.readInt();
76 final byte[] bytes = new byte[bytesLength];
78 final String text = new String(bytes, "UTF8");
80 parseSection(heading, text);
83 if (pageCount % 1000 == 0) {
84 LOG.info("pageCount=" + pageCount);
88 LOG.info("***COUNTERS***");
89 for (final Map.Entry<String, AtomicInteger> entry : counters.entrySet()) {
90 LOG.info(entry.getKey() + ": " + entry.getValue());
95 static final Pattern whitespace = Pattern.compile("\\s+");
96 static String trim(final String s) {
97 return whitespace.matcher(s).replaceAll(" ").trim();
100 public void incrementCount(final String string) {
101 AtomicInteger counter = counters.get(string);
102 if (counter != null) {
103 counter = new AtomicInteger();
104 counters.put(string, counter);
106 counter.incrementAndGet();
110 // -------------------------------------------------------------------------
112 static final class AppendAndIndexWikiCallback<T extends AbstractWiktionaryParser> implements WikiTokenizer.Callback {
115 StringBuilder builder;
116 IndexedEntry indexedEntry;
117 IndexBuilder indexBuilder;
118 final Map<String,FunctionCallback<T>> functionCallbacks = new LinkedHashMap<String, FunctionCallback<T>>();
120 boolean entryTypeNameSticks = false;
121 EntryTypeName entryTypeName = null;
123 final Map<String,AtomicInteger> langCodeToTCount = new LinkedHashMap<String, AtomicInteger>();
125 final NameAndArgs<T> nameAndArgs = new NameAndArgs<T>();
127 public AppendAndIndexWikiCallback(final T parser) {
128 this.parser = parser;
131 public void reset(final StringBuilder builder, final IndexedEntry indexedEntry) {
132 this.builder = builder;
133 this.indexedEntry = indexedEntry;
134 this.indexBuilder = null;
135 entryTypeName = null;
136 entryTypeNameSticks = false;
139 public void dispatch(final String wikiText, final IndexBuilder indexBuilder, final EntryTypeName entryTypeName) {
140 final IndexBuilder oldIndexBuilder = this.indexBuilder;
141 final EntryTypeName oldEntryTypeName = this.entryTypeName;
142 this.indexBuilder = indexBuilder;
143 if (!entryTypeNameSticks) {
144 this.entryTypeName = EnumUtil.min(entryTypeName, this.entryTypeName);
146 if (entryTypeName == null) this.entryTypeName = null;
147 WikiTokenizer.dispatch(wikiText, false, this);
148 this.indexBuilder = oldIndexBuilder;
149 this.entryTypeName = oldEntryTypeName;
152 public void dispatch(final String wikiText, final EntryTypeName entryTypeName) {
153 dispatch(wikiText, this.indexBuilder, entryTypeName);
157 public void onPlainText(final String plainText) {
158 // The only non-recursive callback. Just appends to the builder, and indexes.
159 builder.append(plainText);
160 if (indexBuilder != null && entryTypeName != null && indexedEntry != null) {
161 indexBuilder.addEntryWithString(indexedEntry, plainText, entryTypeName);
166 public void onWikiLink(WikiTokenizer wikiTokenizer) {
167 final String text = wikiTokenizer.wikiLinkText();
168 @SuppressWarnings("unused")
169 final String link = wikiTokenizer.wikiLinkDest();
170 dispatch(text, entryTypeName);
174 public void onFunction(
175 final WikiTokenizer wikiTokenizer,
177 final List<String> args,
178 final Map<String, String> namedArgs) {
180 FunctionCallback<T> functionCallback = functionCallbacks.get(name);
181 if (functionCallback == null || !functionCallback.onWikiFunction(wikiTokenizer, name, args, namedArgs, parser, this)) {
182 // Default function handling:
183 // namedArgs.keySet().removeAll(EnWiktionaryXmlParser.USELESS_WIKI_ARGS);
184 final boolean single = args.isEmpty() && namedArgs.isEmpty();
185 builder.append(single ? "{" : "{{");
187 final IndexBuilder oldIndexBuilder = indexBuilder;
189 nameAndArgs.onWikiFunction(wikiTokenizer, name, args, namedArgs, parser, this);
190 indexBuilder = oldIndexBuilder;
192 builder.append(single ? "}" : "}}");
197 public void onHtml(WikiTokenizer wikiTokenizer) {
198 // Unindexed for now.
199 builder.append(wikiTokenizer.token());
203 public void onMarkup(WikiTokenizer wikiTokenizer) {
208 public final void onComment(WikiTokenizer wikiTokenizer) {
213 public final void onNewline(WikiTokenizer wikiTokenizer) {
218 public final void onHeading(WikiTokenizer wikiTokenizer) {
223 public final void onListItem(WikiTokenizer wikiTokenizer) {
228 // --------------------------------------------------------------------
230 static final class NameAndArgs<T extends AbstractWiktionaryParser> implements FunctionCallback<T> {
232 public boolean onWikiFunction(final WikiTokenizer wikiTokenizer, final String name, final List<String> args,
233 final Map<String, String> namedArgs, final T parser,
234 final AppendAndIndexWikiCallback<T> appendAndIndexWikiCallback) {
237 appendAndIndexWikiCallback.builder.append(name);
239 for (int i = 0; i < args.size(); ++i) {
240 if (args.get(i).length() > 0) {
241 appendAndIndexWikiCallback.builder.append("|");
242 appendAndIndexWikiCallback.dispatch(args.get(i), null, null);
245 appendNamedArgs(namedArgs, appendAndIndexWikiCallback);
249 static NameAndArgs<AbstractWiktionaryParser> NAME_AND_ARGS = new NameAndArgs<AbstractWiktionaryParser>();
251 static void appendNamedArgs(final Map<String, String> namedArgs,
252 final AppendAndIndexWikiCallback<?> appendAndIndexWikiCallback) {
253 for (final Map.Entry<String, String> entry : namedArgs.entrySet()) {
254 appendAndIndexWikiCallback.builder.append("|");
255 appendAndIndexWikiCallback.dispatch(entry.getKey(), null, null);
256 appendAndIndexWikiCallback.builder.append("=");
257 EntryTypeName entryTypeName = null;
258 IndexBuilder indexBuilder = null;
259 // This doesn't work: we'd need to add to word-forms.
260 // System.out.println(entry.getKey());
261 // if (entry.getKey().equals("tr")) {
262 // entryTypeName = EntryTypeName.WIKTIONARY_TRANSLITERATION;
263 // indexBuilder = appendAndIndexWikiCallback.parser.foreignIndexBuilder;
265 appendAndIndexWikiCallback.dispatch(entry.getValue(), indexBuilder, entryTypeName);