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;
17 import java.io.DataInput;
18 import java.io.DataOutput;
20 import java.io.IOException;
21 import java.io.PrintStream;
22 import java.io.RandomAccessFile;
23 import java.nio.BufferUnderflowException;
24 import java.nio.MappedByteBuffer;
25 import java.nio.channels.FileChannel;
26 import java.util.ArrayList;
27 import java.util.Collections;
28 import java.util.List;
30 import com.hughes.android.dictionary.DictionaryInfo;
31 import com.hughes.util.CachingList;
32 import com.hughes.util.DataInputBuffer;
33 import com.hughes.util.raf.RAFList;
34 import com.hughes.util.raf.RAFListSerializer;
36 public class Dictionary {
38 private static final int CACHE_SIZE = 5000;
40 private static final int CURRENT_DICT_VERSION = 7;
41 private static final String END_OF_DICTIONARY = "END OF DICTIONARY";
44 final int dictFileVersion;
45 public final long creationMillis;
46 public final String dictInfo;
47 public final List<PairEntry> pairEntries;
48 public final List<TextEntry> textEntries;
49 public final List<HtmlEntry> htmlEntries;
50 public final List<DataInputBuffer> htmlData;
51 public final List<EntrySource> sources;
52 public final List<Index> indices;
53 // Could be a local variable in constructor, but
54 // this way avoids a native-image VM bug.
55 private final MappedByteBuffer wholefile;
58 * dictFileVersion 1 adds: <li>links to sources? dictFileVersion 2 adds: <li>
59 * counts of tokens in indices.
62 public Dictionary(final String dictInfo) {
63 this.dictFileVersion = CURRENT_DICT_VERSION;
64 this.creationMillis = System.currentTimeMillis();
65 this.dictInfo = dictInfo;
66 pairEntries = new ArrayList<>();
67 textEntries = new ArrayList<>();
68 htmlEntries = new ArrayList<>();
70 sources = new ArrayList<>();
71 indices = new ArrayList<>();
75 public Dictionary(final FileChannel ch) throws IOException {
76 wholefile = ch.map(FileChannel.MapMode.READ_ONLY, 0, ch.size());
77 DataInputBuffer in = new DataInputBuffer(wholefile, 0);
78 dictFileVersion = in.readInt();
79 if (dictFileVersion < 0 || dictFileVersion > CURRENT_DICT_VERSION) {
80 throw new IOException("Invalid dictionary version: " + dictFileVersion);
82 creationMillis = in.readLong();
83 dictInfo = in.readUTF();
85 // Load the sources, then seek past them, because reading them later
86 // disrupts the offset.
88 final RAFList<EntrySource> rafSources = RAFList.create(in, new EntrySource.Serializer(
89 this), dictFileVersion, dictInfo + " sources: ");
90 sources = new ArrayList<>(rafSources);
91 ch.position(rafSources.getEndOffset());
93 pairEntries = CachingList.create(
94 RAFList.create(in, new PairEntry.Serializer(this), dictFileVersion, dictInfo + " pairs: "),
96 textEntries = CachingList.create(
97 RAFList.create(in, new TextEntry.Serializer(this), dictFileVersion, dictInfo + " text: "),
99 if (dictFileVersion >= 5) {
100 htmlEntries = CachingList.create(
101 RAFList.create(in, new HtmlEntry.Serializer(this), dictFileVersion, dictInfo + " html: "),
104 htmlEntries = Collections.emptyList();
106 if (dictFileVersion >= 7) {
107 htmlData = RAFList.create(in, new HtmlEntry.DataDeserializer(), dictFileVersion, dictInfo + " html: ");
111 indices = CachingList.createFullyCached(RAFList.create(in, new IndexSerializer(),
112 dictFileVersion, dictInfo + " index: "));
113 } catch (RuntimeException e) {
114 throw new IOException("RuntimeException loading dictionary", e);
116 final String end = in.readUTF();
117 if (!end.equals(END_OF_DICTIONARY)) {
118 throw new IOException("Dictionary seems corrupt: " + end);
122 public void write(DataOutput out) throws IOException {
123 RandomAccessFile raf = (RandomAccessFile)out;
124 if (dictFileVersion < 7) throw new RuntimeException("write function cannot write formats older than v7!");
125 raf.writeInt(dictFileVersion);
126 raf.writeLong(creationMillis);
127 raf.writeUTF(dictInfo);
128 System.out.println("sources start: " + raf.getFilePointer());
129 RAFList.write(raf, sources, new EntrySource.Serializer(this));
130 System.out.println("pair start: " + raf.getFilePointer());
131 RAFList.write(raf, pairEntries, new PairEntry.Serializer(this), 64, true);
132 System.out.println("text start: " + raf.getFilePointer());
133 RAFList.write(raf, textEntries, new TextEntry.Serializer(this));
134 System.out.println("html index start: " + raf.getFilePointer());
135 RAFList.write(raf, htmlEntries, new HtmlEntry.Serializer(this), 64, true);
136 System.out.println("html data start: " + raf.getFilePointer());
137 assert htmlData == null;
138 RAFList.write(raf, htmlEntries, new HtmlEntry.DataSerializer(), 128, true);
139 System.out.println("indices start: " + raf.getFilePointer());
140 RAFList.write(raf, indices, new IndexSerializer());
141 System.out.println("end: " + raf.getFilePointer());
142 raf.writeUTF(END_OF_DICTIONARY);
145 private final class IndexSerializer implements RAFListSerializer<Index> {
147 public Index read(DataInput raf, final int readIndex) throws IOException {
148 return new Index(Dictionary.this, (DataInputBuffer)raf);
152 public void write(DataOutput raf, Index t) throws IOException {
157 final RAFListSerializer<HtmlEntry> htmlEntryIndexSerializer = new RAFListSerializer<HtmlEntry>() {
159 public void write(DataOutput raf, HtmlEntry t) {
164 public HtmlEntry read(DataInput raf, int readIndex) throws IOException {
165 return htmlEntries.get(raf.readInt());
169 public void print(final PrintStream out) {
170 out.println("dictInfo=" + dictInfo);
171 for (final EntrySource entrySource : sources) {
172 out.printf("EntrySource: %s %d\n", entrySource.name, entrySource.numEntries);
175 for (final Index index : indices) {
176 out.printf("Index: %s %s\n", index.shortName, index.longName);
182 public DictionaryInfo getDictionaryInfo() {
183 final DictionaryInfo result = new DictionaryInfo();
184 result.creationMillis = this.creationMillis;
185 result.dictInfo = this.dictInfo;
186 for (final Index index : indices) {
187 result.indexInfos.add(index.getIndexInfo());
192 // get DictionaryInfo for case when Dictionary cannot be opened
193 private static DictionaryInfo getErrorDictionaryInfo(final File file) {
194 final DictionaryInfo dictionaryInfo = new DictionaryInfo();
195 dictionaryInfo.uncompressedFilename = file.getName();
196 dictionaryInfo.uncompressedBytes = file.length();
197 return dictionaryInfo;
200 public static DictionaryInfo getDictionaryInfo(final File file) {
201 RandomAccessFile raf = null;
203 raf = new RandomAccessFile(file, "r");
204 final Dictionary dict = new Dictionary(raf.getChannel());
205 final DictionaryInfo dictionaryInfo = dict.getDictionaryInfo();
206 dictionaryInfo.uncompressedFilename = file.getName();
207 dictionaryInfo.uncompressedBytes = file.length();
209 return dictionaryInfo;
210 } catch (IOException e) {
211 return getErrorDictionaryInfo(file);
212 } catch (IllegalArgumentException e) {
213 // Most likely due to a Buffer.limit beyond size of file,
214 // do not crash just because of a truncated dictionary file
215 return getErrorDictionaryInfo(file);
216 } catch (BufferUnderflowException e) {
217 // Most likely due to a read beyond the buffer limit set,
218 // do not crash just because of a truncated or corrupt dictionary file
219 return getErrorDictionaryInfo(file);
224 } catch (IOException e) {