puzzle-generator/src/main/java/puzzle/Meta.java

package puzzle;

import com.google.gson.Gson;
import puzzle.SwedishGenerator.Lemma;
import java.io.IOException;
import java.nio.ByteBuffer;
import java.nio.ByteOrder;
import java.nio.MappedByteBuffer;
import java.nio.channels.FileChannel;
import java.nio.charset.StandardCharsets;
import java.nio.file.Path;
import java.nio.file.StandardOpenOption;
import java.util.Arrays;
import java.util.Locale;

public class Meta {
   
   static final Gson GSON    = new Gson();
   static final int  VERSION = 1;
   
   static final int SHARD_MAGIC = 0x49445831; // "IDX1"
   static final int MAP_MAGIC   = 0x4D415031; // "MAP1"
   
   static final ByteOrder ORDER = ByteOrder.BIG_ENDIAN;
   
   static record ShardRec(String word, long w, int simpel, String[] clues) { }
   
   static final Path projectRoot = Path.of("").toAbsolutePath().normalize(); // current working dir
   //static final Path dir         = projectRoot.resolve("src/main/resources/shards");
   //static final Path shardData   = dir.resolve("shard0.data");
   //static final Path shardMap    = dir.resolve("shard0.map");
   static final Path dir       = detectShardDir();
   static final Path shardData = dir.resolve("shard0.data");
   static final Path shardMap  = dir.resolve("shard0.map");
   
   private static Path detectShardDir() {
      // 1) optioneel override
      String p = System.getProperty("puzzle.shards.dir");
      if (p != null && !p.isBlank()) return Path.of(p).toAbsolutePath().normalize();
      
      // 2) default: naast classes output (CLASS_OUTPUT/shards)
      try {
         var url = Meta.class.getProtectionDomain().getCodeSource().getLocation(); // classes dir (niet jar)
         Path classes = Path.of(url.toURI());
         return classes.resolve("shards");
      } catch (Exception e) {
         // 3) fallback: oude dev-locatie
         return Path.of("").toAbsolutePath().normalize().resolve("src/main/resources/shards");
      }
   }
   // --- Lookup: w -> i using mmap ---
   static int findIndexInMapMmap(long target) throws IOException {
      try (var ch = FileChannel.open(Meta.shardMap, StandardOpenOption.READ)) {
         var mbb = (MappedByteBuffer) ch.map(FileChannel.MapMode.READ_ONLY, 0, ch.size()).order(ORDER);
         
         var magic = mbb.getInt(0);
         var ver   = mbb.getInt(4);
         var n     = mbb.getInt(8);
         if (magic != MAP_MAGIC || ver != VERSION) throw new IOException("Bad map file");
         
         int lo = 0, hi = n - 1;
         while (lo <= hi) {
            var mid = (lo + hi) >>> 1;
            var off = 12 + mid * 8;
            var key = mbb.getLong(off);
            
            if (key < target) lo = mid + 1;
            else if (key > target) hi = mid - 1;
            else return mid;
         }
         return -1;
      }
   }
   
   // --- Read record i from shard.data (your format) ---
   static ShardLem readRecord(long w, int i) throws IOException {
      try (var ch = FileChannel.open(Meta.shardData, StandardOpenOption.READ)) {
         var hdr = ByteBuffer.allocate(12).order(ORDER);
         ch.read(hdr);
         hdr.flip();
         var magic = hdr.getInt();
         var ver   = hdr.getInt();
         var n     = hdr.getInt();
         if (magic != SHARD_MAGIC || ver != VERSION) throw new IOException("Bad shard file");
         if (i < 0 || i >= n) throw new IndexOutOfBoundsException();
         
         var tableStart = 12L;
         var dataStart  = 12L + (long) n * 4L;
         
         var offI = readIntAt(ch, tableStart + (long) i * 4L);
         var offIp = (i + 1 < n)
                     ? readIntAt(ch, tableStart + (long) (i + 1) * 4L)
                     : (int) (ch.size() - dataStart);
         
         var len = offIp - offI;
         var buf = ByteBuffer.allocate(len);
         ch.position(dataStart + offI);
         ch.read(buf);
         buf.flip();
         
         var s     = StandardCharsets.UTF_8.decode(buf).toString();
         var parts = s.split("\t", 3);
         
         var simpel = Integer.parseInt(parts[1]);
         var clues  = GSON.fromJson(parts[2], String[].class);
         return new ShardLem(w, simpel, clues);
      }
   }
   
   static int readIntAt(FileChannel ch, long pos) throws IOException {
      var b = ByteBuffer.allocate(4).order(ORDER);
      ch.position(pos);
      ch.read(b);
      b.flip();
      return b.getInt();
   }
   
   // --- Demo main ---
   public static ShardLem lookup(long w) {
      try {
         var i = findIndexInMapMmap(Lemma.pack43(w));
         System.out.println("\nQuery: w=" + w + " -> i=" + i);
         if (i >= 0) {
            var rec = readRecord(w, i);
            System.out.println("  simpel=" + rec.simpel());
            System.out.println("  clues=" + Arrays.toString(rec.clues()));
            return rec;
         } else {
            System.out.println("  NOT FOUND");
            throw new RuntimeException("NOT FOUND");
         }
      } catch (Exception e) {
         throw new RuntimeException(e);
      }
   }
   public record ShardLem(long w, int simpel, String[] clues) { }
}