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.enwiktionary.EnWiktionaryXmlParser;
35 import com.hughes.util.Args;
36 import com.hughes.util.FileUtil;
39 public class DictionaryBuilder {
41 public final Dictionary dictionary;
42 public final List<IndexBuilder> indexBuilders = new ArrayList<IndexBuilder>();
44 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) {
45 dictionary = new Dictionary(dictInfoString);
46 indexBuilders.add(new IndexBuilder(this, lang0.getIsoCode(), lang0.getIsoCode() + "->" + lang1.getIsoCode(), lang0, normalizerRules1, lang1Stoplist, false));
47 indexBuilders.add(new IndexBuilder(this, lang1.getIsoCode(), lang1.getIsoCode() + "->" + lang0.getIsoCode(), lang1, normalizerRules2, lang2Stoplist, true));
51 for (final IndexBuilder indexBuilder : indexBuilders) {
53 dictionary.indices.add(indexBuilder.index);
57 public static void main(final String[] args) throws IOException, ParserConfigurationException, SAXException {
58 System.out.println("Running with arguments:");
59 for (final String arg : args) {
60 System.out.println(arg);
63 final Map<String,String> keyValueArgs = Args.keyValueArgs(args);
65 if (!keyValueArgs.containsKey("lang1") || !keyValueArgs.containsKey("lang2")) {
66 fatalError("--lang1= and --lang2= must both be specified.");
68 final Language lang1 = Language.lookup(keyValueArgs.remove("lang1"));
69 final Language lang2 = Language.lookup(keyValueArgs.remove("lang2"));
71 final Set<String> lang1Stoplist = new LinkedHashSet<String>();
72 final Set<String> lang2Stoplist = new LinkedHashSet<String>();
73 final String lang1StoplistFile = keyValueArgs.remove("lang1Stoplist");
74 final String lang2StoplistFile = keyValueArgs.remove("lang2Stoplist");
75 if (lang1StoplistFile != null) {
76 lang1Stoplist.addAll(FileUtil.readLines(new File(lang1StoplistFile)));
78 if (lang2StoplistFile != null) {
79 lang2Stoplist.addAll(FileUtil.readLines(new File(lang2StoplistFile)));
82 String normalizerRules1 = keyValueArgs.remove("normalizerRules1");
83 String normalizerRules2 = keyValueArgs.remove("normalizerRules2");
84 if (normalizerRules1 == null) {
85 normalizerRules1 = lang1.getDefaultNormalizerRules();
87 if (normalizerRules2 == null) {
88 normalizerRules2 = lang2.getDefaultNormalizerRules();
91 final String dictOutFilename = keyValueArgs.remove("dictOut");
92 if (dictOutFilename == null) {
93 fatalError("--dictOut= must be specified.");
96 String dictInfo = keyValueArgs.remove("dictInfo");
97 if (dictInfo == null) {
98 fatalError("--dictInfo= must be specified.");
100 if (dictInfo.startsWith("@")) {
101 dictInfo = FileUtil.readToString(new File(dictInfo.substring(1)));
104 final String printFile = keyValueArgs.remove("print");
106 System.out.println("lang1=" + lang1);
107 System.out.println("lang2=" + lang2);
108 System.out.println("normalizerRules1=" + normalizerRules1);
109 System.out.println("normalizerRules2=" + normalizerRules2);
110 System.out.println("dictInfo=" + dictInfo);
111 System.out.println("dictOut=" + dictOutFilename);
113 final DictionaryBuilder dictionaryBuilder = new DictionaryBuilder(dictInfo, lang1, lang2, normalizerRules1, normalizerRules2, lang1Stoplist, lang2Stoplist);
115 for (int i = 0; i < 100; ++i) {
116 final String prefix = "input" + i;
117 if (keyValueArgs.containsKey(prefix)) {
118 final File file = new File(keyValueArgs.remove(prefix));
119 System.out.println("Processing: " + file);
120 String charsetName = keyValueArgs.remove(prefix + "Charset");
121 if (charsetName == null) {
122 charsetName = "UTF8";
124 final Charset charset = Charset.forName(charsetName);
125 String inputName = keyValueArgs.remove(prefix + "Name");
126 if (inputName == null) {
127 fatalError("Must specify human readable name for: " + prefix + "Name");
130 final EntrySource entrySource = new EntrySource(dictionaryBuilder.dictionary.sources.size(), inputName, 0);
131 System.out.println("");
133 String inputFormat = keyValueArgs.remove(prefix + "Format");
134 if ("tab_separated".equals(inputFormat)) {
135 final boolean flipColumns = "true".equals(keyValueArgs.remove(prefix + "FlipColumns"));
136 new DictFileParser(charset, flipColumns, DictFileParser.TAB, null, dictionaryBuilder, dictionaryBuilder.indexBuilders.toArray(new IndexBuilder[0]), null).parseFile(file, entrySource);
137 } else if ("chemnitz".equals(inputFormat)) {
138 final boolean flipColumns = "true".equals(keyValueArgs.remove(prefix + "FlipColumns"));
139 new DictFileParser(charset, flipColumns, DictFileParser.DOUBLE_COLON, DictFileParser.PIPE, dictionaryBuilder, dictionaryBuilder.indexBuilders.toArray(new IndexBuilder[0]), null).parseFile(file, entrySource);
140 } else if ("enwiktionary".equals(inputFormat)) {
141 final Pattern langPattern = Pattern.compile(keyValueArgs.remove(prefix + "LangPattern"), Pattern.CASE_INSENSITIVE);
142 final Pattern langCodePattern = Pattern.compile(keyValueArgs.remove(prefix + "LangCodePattern"));
143 final int enIndex = Integer.parseInt(keyValueArgs.remove(prefix + "EnIndex")) - 1;
144 String pageLimit = keyValueArgs.remove(prefix + "PageLimit");
145 if (pageLimit == null) {
149 if (enIndex < 0 || enIndex >= 2) {
150 fatalError("Must be 1 or 2: " + prefix + "EnIndex");
152 new EnWiktionaryXmlParser(dictionaryBuilder.indexBuilders.get(enIndex), dictionaryBuilder.indexBuilders.get(1-enIndex),
153 langPattern, langCodePattern, enIndex != 0).parse(file, entrySource, Integer.parseInt(pageLimit));
155 fatalError("Invalid or missing input format: " + inputFormat);
158 dictionaryBuilder.dictionary.sources.add(entrySource);
159 System.out.println("Done: " + file + "\n\n");
163 dictionaryBuilder.build();
165 if (printFile != null) {
166 final PrintStream out = new PrintStream(new File(printFile));
167 dictionaryBuilder.dictionary.print(out);
171 System.out.println("Writing dictionary to: " + dictOutFilename);
172 final RandomAccessFile dictOut = new RandomAccessFile(dictOutFilename, "rw");
173 dictOut.setLength(0);
174 dictionaryBuilder.dictionary.write(dictOut);
177 if (!keyValueArgs.isEmpty()) {
178 System.err.println("WARNING: couldn't parse arguments: " + keyValueArgs);
184 private static void fatalError(String string) {
185 System.err.println(string);