1 // Copyright 2011 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.engine;
18 import java.io.IOException;
19 import java.io.PrintStream;
20 import java.io.RandomAccessFile;
21 import java.nio.charset.Charset;
22 import java.util.ArrayList;
23 import java.util.LinkedHashSet;
24 import java.util.List;
27 import java.util.regex.Pattern;
29 import javax.xml.parsers.ParserConfigurationException;
31 import org.xml.sax.SAXException;
33 import com.hughes.android.dictionary.parser.DictFileParser;
34 import com.hughes.android.dictionary.parser.Parser;
35 import com.hughes.android.dictionary.parser.wiktionary.EnForeignParser;
36 import com.hughes.android.dictionary.parser.wiktionary.EnToTranslationParser;
37 import com.hughes.android.dictionary.parser.wiktionary.EnTranslationToTranslationParser;
38 import com.hughes.util.Args;
39 import com.hughes.util.FileUtil;
41 public class DictionaryBuilder {
43 public final Dictionary dictionary;
44 public final List<IndexBuilder> indexBuilders = new ArrayList<IndexBuilder>();
46 public DictionaryBuilder(final String dictInfoString, final Language lang0, final Language lang1, final String normalizerRules1, final String normalizerRules2, final Set<String> lang1Stoplist, final Set<String> lang2Stoplist) {
47 dictionary = new Dictionary(dictInfoString);
48 indexBuilders.add(new IndexBuilder(this, lang0.getIsoCode(), lang0.getIsoCode() + "->" + lang1.getIsoCode(), lang0, normalizerRules1, lang1Stoplist, false));
49 indexBuilders.add(new IndexBuilder(this, lang1.getIsoCode(), lang1.getIsoCode() + "->" + lang0.getIsoCode(), lang1, normalizerRules2, lang2Stoplist, true));
53 for (final IndexBuilder indexBuilder : indexBuilders) {
55 dictionary.indices.add(indexBuilder.index);
59 public static void main(final String[] args) throws IOException, ParserConfigurationException, SAXException {
60 System.out.println("Running with arguments:");
61 for (final String arg : args) {
62 System.out.println(arg);
65 final Map<String,String> keyValueArgs = Args.keyValueArgs(args);
67 if (!keyValueArgs.containsKey("lang1") || !keyValueArgs.containsKey("lang2")) {
68 fatalError("--lang1= and --lang2= must both be specified.");
70 final Language lang1 = Language.lookup(keyValueArgs.remove("lang1"));
71 final Language lang2 = Language.lookup(keyValueArgs.remove("lang2"));
73 final Set<String> lang1Stoplist = new LinkedHashSet<String>();
74 final Set<String> lang2Stoplist = new LinkedHashSet<String>();
75 final String lang1StoplistFile = keyValueArgs.remove("lang1Stoplist");
76 final String lang2StoplistFile = keyValueArgs.remove("lang2Stoplist");
77 if (lang1StoplistFile != null) {
78 lang1Stoplist.addAll(FileUtil.readLines(new File(lang1StoplistFile)));
80 if (lang2StoplistFile != null) {
81 lang2Stoplist.addAll(FileUtil.readLines(new File(lang2StoplistFile)));
84 String normalizerRules1 = keyValueArgs.remove("normalizerRules1");
85 String normalizerRules2 = keyValueArgs.remove("normalizerRules2");
86 if (normalizerRules1 == null) {
87 normalizerRules1 = lang1.getDefaultNormalizerRules();
89 if (normalizerRules2 == null) {
90 normalizerRules2 = lang2.getDefaultNormalizerRules();
93 final String dictOutFilename = keyValueArgs.remove("dictOut");
94 if (dictOutFilename == null) {
95 fatalError("--dictOut= must be specified.");
98 String dictInfo = keyValueArgs.remove("dictInfo");
99 if (dictInfo == null) {
100 fatalError("--dictInfo= must be specified.");
102 if (dictInfo.startsWith("@")) {
103 dictInfo = FileUtil.readToString(new File(dictInfo.substring(1)));
106 final String printFile = keyValueArgs.remove("print");
108 System.out.println("lang1=" + lang1);
109 System.out.println("lang2=" + lang2);
110 System.out.println("normalizerRules1=" + normalizerRules1);
111 System.out.println("normalizerRules2=" + normalizerRules2);
112 System.out.println("dictInfo=" + dictInfo);
113 System.out.println("dictOut=" + dictOutFilename);
115 final DictionaryBuilder dictionaryBuilder = new DictionaryBuilder(dictInfo, lang1, lang2, normalizerRules1, normalizerRules2, lang1Stoplist, lang2Stoplist);
117 for (int i = 0; i < 100; ++i) {
118 final String prefix = "input" + i;
119 if (keyValueArgs.containsKey(prefix)) {
120 final File file = new File(keyValueArgs.remove(prefix));
121 System.out.println("Processing: " + file);
122 String charsetName = keyValueArgs.remove(prefix + "Charset");
123 if (charsetName == null) {
124 charsetName = "UTF8";
126 final Charset charset = Charset.forName(charsetName);
127 String inputName = keyValueArgs.remove(prefix + "Name");
128 if (inputName == null) {
129 fatalError("Must specify human readable name for: " + prefix + "Name");
131 String pageLimitString = keyValueArgs.remove(prefix + "PageLimit");
132 if (pageLimitString == null) {
133 pageLimitString = "-1";
135 final int pageLimit = Integer.parseInt(pageLimitString);
137 final EntrySource entrySource = new EntrySource(dictionaryBuilder.dictionary.sources.size(), inputName, 0);
138 System.out.println("");
140 String inputFormat = keyValueArgs.remove(prefix + "Format");
141 if ("tab_separated".equals(inputFormat)) {
142 final boolean flipColumns = "true".equals(keyValueArgs.remove(prefix + "FlipColumns"));
143 new DictFileParser(charset, flipColumns, DictFileParser.TAB, null, dictionaryBuilder, dictionaryBuilder.indexBuilders.toArray(new IndexBuilder[0]), null).parse(file, entrySource, pageLimit);
144 } else if ("chemnitz".equals(inputFormat)) {
145 final boolean flipColumns = "true".equals(keyValueArgs.remove(prefix + "FlipColumns"));
146 new DictFileParser(charset, flipColumns, DictFileParser.DOUBLE_COLON, DictFileParser.PIPE, dictionaryBuilder, dictionaryBuilder.indexBuilders.toArray(new IndexBuilder[0]), null).parse(file, entrySource, pageLimit);
147 } else if ("enwiktionary".equals(inputFormat)) {
148 final String type = keyValueArgs.remove(prefix + "WiktionaryType");
149 final Pattern langPattern = Pattern.compile(keyValueArgs.remove(prefix + "LangPattern"), Pattern.CASE_INSENSITIVE);
150 final Pattern langCodePattern = Pattern.compile(keyValueArgs.remove(prefix + "LangCodePattern"));
151 final int enIndex = Integer.parseInt(keyValueArgs.remove(prefix + "EnIndex")) - 1;
153 if (enIndex < 0 || enIndex >= 2) {
154 fatalError("Must be 1 or 2: " + prefix + "EnIndex");
157 if ("EnToTranslation".equals(type)) {
158 parser = new EnToTranslationParser(dictionaryBuilder.indexBuilders.get(enIndex), dictionaryBuilder.indexBuilders.get(1-enIndex),
159 langPattern, langCodePattern, enIndex != 0);
160 } else if ("EnForeign".equals(type)) {
161 parser = new EnForeignParser(dictionaryBuilder.indexBuilders.get(enIndex), dictionaryBuilder.indexBuilders.get(1-enIndex),
162 langPattern, langCodePattern, enIndex != 0);
164 fatalError("Invalid WiktionaryType (use EnToTranslation or EnForeign): " + type);
167 parser.parse(file, entrySource, pageLimit);
168 } else if (EnTranslationToTranslationParser.NAME.equals(inputFormat)) {
169 final Pattern codePattern1 = Pattern.compile(keyValueArgs.remove(prefix + "LangPattern1"), Pattern.CASE_INSENSITIVE);
170 final Pattern codePattern2 = Pattern.compile(keyValueArgs.remove(prefix + "LangPattern2"), Pattern.CASE_INSENSITIVE);
171 new EnTranslationToTranslationParser(dictionaryBuilder.indexBuilders, new Pattern[] {codePattern1, codePattern2}).parse(file, entrySource, pageLimit);
173 fatalError("Invalid or missing input format: " + inputFormat);
176 dictionaryBuilder.dictionary.sources.add(entrySource);
177 System.out.println("Done: " + file + "\n\n");
181 dictionaryBuilder.build();
183 if (printFile != null) {
184 final PrintStream out = new PrintStream(new File(printFile));
185 dictionaryBuilder.dictionary.print(out);
189 System.out.println("Writing dictionary to: " + dictOutFilename);
190 final RandomAccessFile dictOut = new RandomAccessFile(dictOutFilename, "rw");
191 dictOut.setLength(0);
192 dictionaryBuilder.dictionary.write(dictOut);
195 if (!keyValueArgs.isEmpty()) {
196 System.err.println("WARNING: couldn't parse arguments: " + keyValueArgs);
202 private static void fatalError(String string) {
203 System.err.println(string);