-import com.hughes.util.raf.RAFSerializable;
-
-
-public class Dictionary implements RAFSerializable<Dictionary> {
-
- // persisted
- final String dictInfo;
- final List<PairEntry> pairEntries;
- final List<TextEntry> textEntries;
- final List<EntrySource> sources;
- final List<Index> indices;
-
- public Dictionary(final String dictInfo) {
- this.dictInfo = dictInfo;
- pairEntries = new ArrayList<PairEntry>();
- textEntries = new ArrayList<TextEntry>();
- sources = new ArrayList<EntrySource>();
- indices = new ArrayList<Index>();
- }
-
- public Dictionary(final RandomAccessFile raf) throws IOException {
- dictInfo = raf.readUTF();
-
- sources = RAFList.create(raf, EntrySource.SERIALIZER, raf.getFilePointer());
-
- // TODO: caching
- pairEntries = RAFList.create(raf, PairEntry.SERIALIZER, raf.getFilePointer());
-
- // TODO: caching
- textEntries = RAFList.create(raf, TextEntry.SERIALIZER, raf.getFilePointer());
-
- final List<Index> rawIndices = RAFList.create(raf, indexSerializer, raf.getFilePointer());
- indices = CachingList.create(rawIndices, rawIndices.size());
- }
-
- public void print(final PrintStream out) {
- out.println("dictInfo=" + dictInfo);
- for (final Index index : indices) {
- index.print(out);
- out.println();
+
+public class 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;
+ 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>
+ * counts of tokens in indices.
+ */
+
+ public Dictionary(final String dictInfo) {
+ this.dictFileVersion = CURRENT_DICT_VERSION;
+ this.creationMillis = System.currentTimeMillis();
+ this.dictInfo = dictInfo;
+ pairEntries = new ArrayList<>();
+ textEntries = new ArrayList<>();
+ htmlEntries = new ArrayList<>();
+ htmlData = null;
+ sources = new ArrayList<>();
+ indices = new ArrayList<>();
+ wholefile = null;
+ }
+
+ 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 = 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(in, new EntrySource.Serializer(
+ this), dictFileVersion, dictInfo + " sources: ");
+ sources = new ArrayList<>(rafSources);
+ ch.position(rafSources.getEndOffset());
+
+ pairEntries = CachingList.create(
+ RAFList.create(in, new PairEntry.Serializer(this), dictFileVersion, dictInfo + " pairs: "),
+ CACHE_SIZE, false);
+ textEntries = CachingList.create(
+ RAFList.create(in, new TextEntry.Serializer(this), dictFileVersion, dictInfo + " text: "),
+ CACHE_SIZE, true);
+ if (dictFileVersion >= 5) {
+ htmlEntries = CachingList.create(
+ RAFList.create(in, new HtmlEntry.Serializer(this), dictFileVersion, dictInfo + " html: "),
+ CACHE_SIZE, false);
+ } else {
+ htmlEntries = Collections.emptyList();
+ }
+ 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) {
+ throw new IOException("RuntimeException loading dictionary", e);
+ }
+ final String end = in.readUTF();
+ if (!end.equals(END_OF_DICTIONARY)) {
+ throw new IOException("Dictionary seems corrupt: " + end);
+ }
+ }
+
+ 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);
+ System.out.println("sources start: " + raf.getFilePointer());
+ RAFList.write(raf, sources, new EntrySource.Serializer(this));
+ System.out.println("pair start: " + raf.getFilePointer());
+ 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 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, new IndexSerializer());
+ System.out.println("end: " + raf.getFilePointer());
+ raf.writeUTF(END_OF_DICTIONARY);