package com.hughes.android.dictionary.engine;
-import com.hughes.android.dictionary.DictionaryInfo;
-import com.hughes.util.CachingList;
-import com.hughes.util.StringUtil;
-import com.hughes.util.raf.RAFList;
-import com.hughes.util.raf.RAFListSerializer;
-import com.hughes.util.raf.RAFSerializable;
-
import java.io.DataInput;
import java.io.DataOutput;
import java.io.File;
import java.io.IOException;
import java.io.PrintStream;
import java.io.RandomAccessFile;
+import java.nio.MappedByteBuffer;
+import java.nio.channels.FileChannel;
import java.util.ArrayList;
import java.util.Collections;
import java.util.List;
-public class Dictionary implements RAFSerializable<Dictionary> {
+import com.hughes.android.dictionary.DictionaryInfo;
+import com.hughes.util.CachingList;
+import com.hughes.util.DataInputBuffer;
+import com.hughes.util.raf.RAFList;
+import com.hughes.util.raf.RAFListSerializer;
- static final int CACHE_SIZE = 5000;
+public class Dictionary {
- static final int CURRENT_DICT_VERSION = 7;
- static final String END_OF_DICTIONARY = "END OF DICTIONARY";
+ private static final int CACHE_SIZE = 5000;
+
+ private static final int CURRENT_DICT_VERSION = 7;
+ private static final String END_OF_DICTIONARY = "END OF DICTIONARY";
// persisted
final int dictFileVersion;
- final long creationMillis;
+ public final long creationMillis;
public final String dictInfo;
public final List<PairEntry> pairEntries;
public final List<TextEntry> textEntries;
public final List<HtmlEntry> htmlEntries;
+ public final List<DataInputBuffer> htmlData;
public final List<EntrySource> sources;
public final List<Index> indices;
+ // Could be a local variable in constructor, but
+ // this way avoids a native-image VM bug.
+ private final MappedByteBuffer wholefile;
/**
* dictFileVersion 1 adds: <li>links to sources? dictFileVersion 2 adds: <li>
this.dictFileVersion = CURRENT_DICT_VERSION;
this.creationMillis = System.currentTimeMillis();
this.dictInfo = dictInfo;
- pairEntries = new ArrayList<PairEntry>();
- textEntries = new ArrayList<TextEntry>();
- htmlEntries = new ArrayList<HtmlEntry>();
- sources = new ArrayList<EntrySource>();
- indices = new ArrayList<Index>();
+ pairEntries = new ArrayList<>();
+ textEntries = new ArrayList<>();
+ htmlEntries = new ArrayList<>();
+ htmlData = null;
+ sources = new ArrayList<>();
+ indices = new ArrayList<>();
+ wholefile = null;
}
- public Dictionary(final RandomAccessFile raf) throws IOException {
- dictFileVersion = raf.readInt();
+ public Dictionary(final FileChannel ch) throws IOException {
+ wholefile = ch.map(FileChannel.MapMode.READ_ONLY, 0, ch.size());
+ DataInputBuffer in = new DataInputBuffer(wholefile, 0);
+ dictFileVersion = in.readInt();
if (dictFileVersion < 0 || dictFileVersion > CURRENT_DICT_VERSION) {
throw new IOException("Invalid dictionary version: " + dictFileVersion);
}
- creationMillis = raf.readLong();
- dictInfo = raf.readUTF();
+ creationMillis = in.readLong();
+ dictInfo = in.readUTF();
// Load the sources, then seek past them, because reading them later
// disrupts the offset.
try {
- final RAFList<EntrySource> rafSources = RAFList.create(raf, new EntrySource.Serializer(
- this), raf.getFilePointer(), dictFileVersion);
- sources = new ArrayList<EntrySource>(rafSources);
- raf.seek(rafSources.getEndOffset());
+ final RAFList<EntrySource> rafSources = RAFList.create(in, new EntrySource.Serializer(
+ this), dictFileVersion, dictInfo + " sources: ");
+ sources = new ArrayList<>(rafSources);
+ ch.position(rafSources.getEndOffset());
pairEntries = CachingList.create(
- RAFList.create(raf, new PairEntry.Serializer(this), raf.getFilePointer(), dictFileVersion, dictFileVersion >= 7 ? 64 : 1, dictFileVersion >= 7),
- CACHE_SIZE);
+ RAFList.create(in, new PairEntry.Serializer(this), dictFileVersion, dictInfo + " pairs: "),
+ CACHE_SIZE, false);
textEntries = CachingList.create(
- RAFList.create(raf, new TextEntry.Serializer(this), raf.getFilePointer(), dictFileVersion),
- CACHE_SIZE);
+ RAFList.create(in, new TextEntry.Serializer(this), dictFileVersion, dictInfo + " text: "),
+ CACHE_SIZE, true);
if (dictFileVersion >= 5) {
htmlEntries = CachingList.create(
- RAFList.create(raf, new HtmlEntry.Serializer(this), raf.getFilePointer(), dictFileVersion),
- CACHE_SIZE);
+ RAFList.create(in, new HtmlEntry.Serializer(this), dictFileVersion, dictInfo + " html: "),
+ CACHE_SIZE, false);
} else {
htmlEntries = Collections.emptyList();
}
- indices = CachingList.createFullyCached(RAFList.create(raf, indexSerializer,
- raf.getFilePointer(), dictFileVersion));
+ if (dictFileVersion >= 7) {
+ htmlData = RAFList.create(in, new HtmlEntry.DataDeserializer(), dictFileVersion, dictInfo + " html: ");
+ } else {
+ htmlData = null;
+ }
+ indices = CachingList.createFullyCached(RAFList.create(in, new IndexSerializer(),
+ dictFileVersion, dictInfo + " index: "));
} catch (RuntimeException e) {
- final IOException ioe = new IOException("RuntimeException loading dictionary");
- ioe.initCause(e);
- throw ioe;
+ throw new IOException("RuntimeException loading dictionary", e);
}
- final String end = raf.readUTF();
+ final String end = in.readUTF();
if (!end.equals(END_OF_DICTIONARY)) {
throw new IOException("Dictionary seems corrupt: " + end);
}
}
- @Override
public void write(DataOutput out) throws IOException {
RandomAccessFile raf = (RandomAccessFile)out;
+ if (dictFileVersion < 7) throw new RuntimeException("write function cannot write formats older than v7!");
raf.writeInt(dictFileVersion);
raf.writeLong(creationMillis);
raf.writeUTF(dictInfo);
RAFList.write(raf, pairEntries, new PairEntry.Serializer(this), 64, true);
System.out.println("text start: " + raf.getFilePointer());
RAFList.write(raf, textEntries, new TextEntry.Serializer(this));
- System.out.println("html start: " + raf.getFilePointer());
- RAFList.write(raf, htmlEntries, new HtmlEntry.Serializer(this));
+ System.out.println("html index start: " + raf.getFilePointer());
+ RAFList.write(raf, htmlEntries, new HtmlEntry.Serializer(this), 64, true);
+ System.out.println("html data start: " + raf.getFilePointer());
+ assert htmlData == null;
+ RAFList.write(raf, htmlEntries, new HtmlEntry.DataSerializer(), 128, true);
System.out.println("indices start: " + raf.getFilePointer());
- RAFList.write(raf, indices, indexSerializer);
+ RAFList.write(raf, indices, new IndexSerializer());
System.out.println("end: " + raf.getFilePointer());
raf.writeUTF(END_OF_DICTIONARY);
}
- private final RAFListSerializer<Index> indexSerializer = new RAFListSerializer<Index>() {
+ private final class IndexSerializer implements RAFListSerializer<Index> {
@Override
public Index read(DataInput raf, final int readIndex) throws IOException {
- return new Index(Dictionary.this, raf);
+ return new Index(Dictionary.this, (DataInputBuffer)raf);
}
@Override
public void write(DataOutput raf, Index t) throws IOException {
t.write(raf);
}
- };
+ }
final RAFListSerializer<HtmlEntry> htmlEntryIndexSerializer = new RAFListSerializer<HtmlEntry>() {
@Override
- public void write(DataOutput raf, HtmlEntry t) throws IOException {
- if (t.index() == -1)
- throw new IndexOutOfBoundsException();
- raf.writeInt(t.index());
+ public void write(DataOutput raf, HtmlEntry t) {
+ assert false;
}
@Override
RandomAccessFile raf = null;
try {
raf = new RandomAccessFile(file, "r");
- final Dictionary dict = new Dictionary(raf);
+ final Dictionary dict = new Dictionary(raf.getChannel());
final DictionaryInfo dictionaryInfo = dict.getDictionaryInfo();
dictionaryInfo.uncompressedFilename = file.getName();
dictionaryInfo.uncompressedBytes = file.length();
raf.close();
return dictionaryInfo;
} catch (IOException e) {
- return null;
+ final DictionaryInfo dictionaryInfo = new DictionaryInfo();
+ dictionaryInfo.uncompressedFilename = file.getName();
+ dictionaryInfo.uncompressedBytes = file.length();
+ return dictionaryInfo;
} finally {
if (raf != null) {
try {
projects / Dictionary.git / blobdiff