allow no config for BytesCodec

Author: brokkoli71

src/main/java/dev/zarr/zarrjava/core/codec/core/BytesCodec.java

@@ -1,38 +1,44 @@
 package dev.zarr.zarrjava.core.codec.core;
 
 import com.fasterxml.jackson.annotation.JsonValue;
+import dev.zarr.zarrjava.ZarrException;
 import dev.zarr.zarrjava.core.codec.ArrayBytesCodec;
 import ucar.ma2.*;
 
 import java.nio.ByteBuffer;
 import java.nio.ByteOrder;
 
 public abstract class BytesCodec extends ArrayBytesCodec {
-    protected abstract ByteOrder getByteOrder();
+    protected abstract ByteOrder getByteOrder() throws ZarrException;
 
     @Override
-public Array decode(ByteBuffer chunkBytes) {
-    chunkBytes.order(getByteOrder());
-    DataType dtype = arrayMetadata.dataType.getMA2DataType();
-    int[] shape = arrayMetadata.chunkShape;
-
-    // Array.factory does not support boolean arrays directly from ByteBuffer
-    if (dtype == DataType.BOOLEAN) {
-        int size = chunkBytes.remaining();
-        boolean[] bools = new boolean[size];
-        for (int i = 0; i < size; i++) {
-            bools[i] = chunkBytes.get(i) != 0;
-        }
+    public Array decode(ByteBuffer chunkBytes) throws ZarrException {
+        ByteOrder order = ByteOrder.BIG_ENDIAN; // Default for 1-byte types
+        if (arrayMetadata.dataType.getByteCount() > 1)
+            order = getByteOrder();
+        chunkBytes.order(order);
+        DataType dtype = arrayMetadata.dataType.getMA2DataType();
+        int[] shape = arrayMetadata.chunkShape;
+
+        // Array.factory does not support boolean arrays directly from ByteBuffer
+        if (dtype == DataType.BOOLEAN) {
+            int size = chunkBytes.remaining();
+            boolean[] bools = new boolean[size];
+            for (int i = 0; i < size; i++) {
+                bools[i] = chunkBytes.get(i) != 0;
+            }
 
-        Index index = Index.factory(shape);
-        return Array.factory(DataType.BOOLEAN, index, bools);
+            Index index = Index.factory(shape);
+            return Array.factory(DataType.BOOLEAN, index, bools);
+        }
+        return Array.factory(dtype, shape, chunkBytes);
     }
 
-    return Array.factory(dtype, shape, chunkBytes);
-}
     @Override
-    public ByteBuffer encode(Array chunkArray) {
-        ByteOrder order = getByteOrder();
+    public ByteBuffer encode(Array chunkArray) throws ZarrException {
+        ByteOrder order = ByteOrder.BIG_ENDIAN; // Default for 1-byte types
+        if (arrayMetadata.dataType.getByteCount() > 1)
+            order = getByteOrder();
 
         // Boolean
         if (chunkArray instanceof ArrayBoolean) {
@@ -94,6 +100,10 @@ public ByteOrder getByteOrder() {
                     throw new RuntimeException("Unreachable");
             }
         }
+
+        public static Endian nativeOrder() {
+            return ByteOrder.nativeOrder() == ByteOrder.LITTLE_ENDIAN ? LITTLE : BIG;
+        }
     }
 
 }

src/main/java/dev/zarr/zarrjava/v3/codec/CodecBuilder.java

@@ -67,14 +67,21 @@ public CodecBuilder withTranspose(int[] order) {
   }
 
   public CodecBuilder withBytes(Endian endian) {
-    codecs.add(new BytesCodec(new BytesCodec.Configuration(endian)));
+    if (dataType.getByteCount() <= 1)
+        codecs.add(new BytesCodec());
+    else
+        codecs.add(new BytesCodec(endian));
     return this;
   }
 
   public CodecBuilder withBytes(String endian) {
     return withBytes(BytesCodec.Endian.valueOf(endian));
   }
 
+  public CodecBuilder withBytes() {
+    return withBytes(Endian.nativeOrder());
+  }
+
   public CodecBuilder withGzip(int clevel) {
     try {
       codecs.add(new GzipCodec(new GzipCodec.Configuration(clevel)));

src/main/java/dev/zarr/zarrjava/v3/codec/core/BytesCodec.java

@@ -1,27 +1,34 @@
 package dev.zarr.zarrjava.v3.codec.core;
 
 import com.fasterxml.jackson.annotation.JsonCreator;
+import com.fasterxml.jackson.annotation.JsonInclude;
 import com.fasterxml.jackson.annotation.JsonProperty;
 import dev.zarr.zarrjava.ZarrException;
 import dev.zarr.zarrjava.v3.codec.Codec;
 import dev.zarr.zarrjava.v3.ArrayMetadata;
 
 import java.nio.ByteOrder;
 import javax.annotation.Nonnull;
+import javax.annotation.Nullable;
 
+@JsonInclude(JsonInclude.Include.NON_NULL)
 public class BytesCodec extends dev.zarr.zarrjava.core.codec.core.BytesCodec implements Codec {
 
   public final String name = "bytes";
-  @Nonnull
+  @Nullable
   public final Configuration configuration;
 
   @JsonCreator
   public BytesCodec(
-      @Nonnull @JsonProperty(value = "configuration", required = true) Configuration configuration
+      @JsonProperty(value = "configuration") Configuration configuration
   ) {
     this.configuration = configuration;
   }
 
+  public BytesCodec() {
+    this((Configuration) null);
+  }
+
   public BytesCodec(Endian endian) {
     this(new BytesCodec.Configuration(endian));
   }
@@ -33,7 +40,10 @@ public long computeEncodedSize(long inputByteLength,
   }
 
   @Override
-  protected ByteOrder getByteOrder() {
+  protected ByteOrder getByteOrder() throws ZarrException {
+    if (configuration == null) {
+      throw new ZarrException("BytesCodec configuration is required to determine endianess.");
+    }
     return configuration.endian.getByteOrder();
   }
 

src/test/java/dev/zarr/zarrjava/ZarrPythonTests.java

@@ -150,9 +150,9 @@ static void assertIsTestdata(ucar.ma2.Array result, dev.zarr.zarrjava.core.DataT
 
     static Stream<Object[]> compressorAndDataTypeProviderV3() {
         Stream<Object[]> datatypeTests = Stream.of(
-//            DataType.BOOL,
-//            DataType.INT8,
-//            DataType.UINT8, // -> BUG: see https://github.com/zarr-developers/zarr-java/issues/27
+            DataType.BOOL,
+            DataType.INT8,
+            DataType.UINT8,
             DataType.INT16,
             DataType.UINT16,
             DataType.INT32,
@@ -203,7 +203,7 @@ public void testReadV3(String codec, String codecParam, DataType dataType) throw
         Assertions.assertArrayEquals(new int[]{2, 4, 8}, array.metadata().chunkShape());
         Assertions.assertEquals(42, array.metadata().attributes.get("answer"));
 
-        assertIsTestdata(result, DataType.INT32);
+        assertIsTestdata(result, dataType);
     }
 
     @ParameterizedTest
@@ -266,7 +266,7 @@ public void testWriteV3(String codec, String codecParam, DataType dataType) thro
         Assertions.assertArrayEquals(new int[]{2, 4, 8}, readArray.metadata().chunkShape());
         Assertions.assertEquals("test_value", readArray.metadata().attributes.get("test_key"));
 
-        assertIsTestdata(result, DataType.INT32);
+        assertIsTestdata(result, dataType);
 
         //read in zarr_python
         run_python_script("zarr_python_read.py", codec, codecParam, dataType.name().toLowerCase(), storeHandle.toPath().toString());

src/test/java/dev/zarr/zarrjava/ZarrV3Test.java

@@ -1,5 +1,6 @@
 package dev.zarr.zarrjava;
 
+import dev.zarr.zarrjava.v3.codec.Codec;
 import dev.zarr.zarrjava.v3.codec.core.BloscCodec;
 import dev.zarr.zarrjava.v3.codec.core.ShardingIndexedCodec;
 
@@ -569,6 +570,20 @@ public void testCreateGroup() throws ZarrException, IOException {
         Assertions.assertEquals("world", group.metadata.attributes.get("hello"));
     }
 
-
-
+    @Test
+    public void testCodecWithoutConfiguration() throws ZarrException, IOException {
+        StoreHandle storeHandle = new FilesystemStore(TESTOUTPUT).resolve("testCodecWithoutConfigurationV3");
+        Array array = Array.create(storeHandle, Array.metadataBuilder()
+            .withShape(10, 10)
+            .withDataType(DataType.UINT8)
+            .withChunkShape(5, 5)
+            .withCodecs(CodecBuilder::withBytes)
+            .build()
+        );
+        Assertions.assertTrue(storeHandle.resolve("zarr.json").exists());
+        Codec bytesCodec = array.metadata().codecs[0];
+        Assertions.assertInstanceOf(BytesCodec.class, bytesCodec);
+        Assertions.assertNull(((BytesCodec) bytesCodec).configuration);
+        // further todo: remove redundant "name" attribute from codec metadata serialization
+    }
 }