MarginaliaSearch/code/common/model/java/nu/marginalia/util/StringPool.java

71 lines
1.6 KiB
Java
Raw Normal View History

package nu.marginalia.util;
2023-01-30 08:40:29 +00:00
import it.unimi.dsi.fastutil.objects.Object2LongOpenHashMap;
import java.util.Arrays;
import java.util.HashMap;
2023-06-12 15:42:28 +00:00
import java.util.Objects;
public class StringPool {
2023-01-30 08:40:29 +00:00
private final HashMap<String, String> words;
2023-01-30 08:40:29 +00:00
private final Object2LongOpenHashMap<String> ages;
private final int maxCap;
2023-01-30 08:40:29 +00:00
long idx;
private StringPool(int capacity, int maxCap) {
this.ages = new Object2LongOpenHashMap<>(capacity);
this.words = new HashMap<>(capacity);
this.maxCap = maxCap;
}
2023-01-30 08:40:29 +00:00
public static StringPool create(int capacity) {
return new StringPool(capacity, capacity * 10);
}
public String internalize(String str) {
2023-01-30 08:40:29 +00:00
prune();
final String ret = words.putIfAbsent(str, str);
2023-01-30 08:40:29 +00:00
ages.put(ret, idx++);
2023-06-12 15:42:28 +00:00
return Objects.requireNonNullElse(ret, str);
}
2023-01-30 08:40:29 +00:00
public String[] internalize(String[] str) {
for (int i = 0; i < str.length; i++) {
str[i] = internalize(str[i]);
}
return str;
}
public void prune() {
if (words.size() < maxCap)
return;
long[] ageValues = ages.values().toLongArray();
Arrays.sort(ageValues);
long cutoff = ageValues[ageValues.length - maxCap / 10];
words.clear();
ages.forEach((word, cnt) -> {
if (cnt >= cutoff) {
words.put(word, word);
}
});
ages.clear();
words.forEach((w,w2) -> {
ages.put(w, idx);
});
}
public void flush() {
words.clear();
}
}