101 lines
3.3 KiB
Java
101 lines
3.3 KiB
Java
package puzzle;
|
|
|
|
import module java.base;
|
|
import lombok.Getter;
|
|
import lombok.NoArgsConstructor;
|
|
import lombok.experimental.Accessors;
|
|
import lombok.val;
|
|
import puzzle.SwedishGenerator.Dict;
|
|
import puzzle.SwedishGenerator.DictEntry;
|
|
import puzzle.SwedishGenerator.Lemma;
|
|
import static java.nio.charset.StandardCharsets.US_ASCII;
|
|
|
|
public final class DictJavaGeneratorMulti {
|
|
|
|
|
|
interface Dicts {
|
|
|
|
static Dict makeDict(long[] wordz) {
|
|
var index = new DictEntryDTO[SwedishGenerator.MAX_WORD_LENGTH_PLUS_ONE];
|
|
Arrays.setAll(index, DictEntryDTO::new);
|
|
for (var lemma : wordz) {
|
|
var L = Lemma.unpackSize(lemma) + 1;//Lemma.unpackSize(lemma) + 2;
|
|
val entry = index[L];
|
|
val idx = entry.words().size();
|
|
val pos = entry.pos();
|
|
entry.words().add(lemma);
|
|
var i = 0;
|
|
for (var w = lemma & Lemma.LETTER_MASK; w != 0; w >>>= 5, i++) {
|
|
pos[i][(int) ((w & 31) - 1)].add(idx);
|
|
}
|
|
}
|
|
for (var i = 2; i < index.length; i++) if (index[i].words().size() <= 0) throw new RuntimeException("No words for length " + i);
|
|
return new Dict(Arrays.stream(index).map(i -> {
|
|
var words = i.words().toArray();
|
|
var numWords = words.length;
|
|
var numLongs = (numWords + 63) >>> 6;
|
|
var bitsets = new long[i.pos().length * 26][numLongs];
|
|
for (var p = 0; p < i.pos().length; p++) {
|
|
for (var l = 0; l < 26; l++) {
|
|
var list = i.pos()[p][l];
|
|
var bs = bitsets[p * 26 + l];
|
|
for (var k = 0; k < list.size(); k++) {
|
|
var wordIdx = list.data()[k];
|
|
bs[wordIdx >>> 6] |= (1L << (wordIdx & 63));
|
|
}
|
|
}
|
|
}
|
|
return new DictEntry(words, bitsets, words.length, (words.length + 63) >>> 6);
|
|
}).toArray(DictEntry[]::new),
|
|
Arrays.stream(index).mapToInt(i -> i.words().size()).sum());
|
|
}
|
|
}
|
|
|
|
|
|
record DictEntryDTO(LongArrayList words, IntListDTO[][] pos) {
|
|
|
|
public DictEntryDTO(int L) {
|
|
this(new LongArrayList(1024), new IntListDTO[L][26]);
|
|
for (var i = 0; i < L; i++) for (var j = 0; j < 26; j++) pos[i][j] = new IntListDTO();
|
|
}
|
|
@Getter
|
|
@Accessors(fluent = true)
|
|
@NoArgsConstructor
|
|
static final class IntListDTO {
|
|
|
|
int[] data = new int[8];
|
|
int size = 0;
|
|
void add(int v) {
|
|
if (size >= data.length) data = Arrays.copyOf(data, data.length * 2);
|
|
data[size++] = v;
|
|
}
|
|
}
|
|
}
|
|
|
|
static final class LongArrayList {
|
|
|
|
long[] a;
|
|
int size;
|
|
|
|
LongArrayList(int initialCapacity) {
|
|
if (initialCapacity < 0) throw new IllegalArgumentException();
|
|
a = new long[initialCapacity];
|
|
}
|
|
|
|
int size() { return size; }
|
|
|
|
void add(long v) {
|
|
if (size == a.length) grow();
|
|
a[size++] = v;
|
|
}
|
|
|
|
void grow() {
|
|
var newCap = a.length == 0 ? 1 : a.length * 2;
|
|
var n = new long[newCap];
|
|
System.arraycopy(a, 0, n, 0, size);
|
|
a = n;
|
|
}
|
|
long[] toArray() { return Arrays.copyOf(a, this.size); }
|
|
}
|
|
}
|