Memory management refactored

kmath-torch/src/nativeMain/kotlin/kscience/kmath/torch/TorchTensor.kt

@@ -15,13 +15,20 @@ public abstract class TorchTensor<T, out TorchTensorBufferImpl : TorchTensorBuff
             )!!
             return TorchTensorFloat(populateStridesFromNative(tensorHandle, rawShape = shape), scope, tensorHandle)
         }
+
         public fun copyFromIntArray(scope: DeferScope, array: IntArray, shape: IntArray): TorchTensorInt {
             val tensorHandle: COpaquePointer = copy_from_blob_int(
                 array.toCValues(), shape.toCValues(), shape.size
             )!!
             return TorchTensorInt(populateStridesFromNative(tensorHandle, rawShape = shape), scope, tensorHandle)
         }
-        public fun copyFromFloatArrayToGPU(scope: DeferScope, array: FloatArray, shape: IntArray, device: Int): TorchTensorFloatGPU {
+
+        public fun copyFromFloatArrayToGPU(
+            scope: DeferScope,
+            array: FloatArray,
+            shape: IntArray,
+            device: Int
+        ): TorchTensorFloatGPU {
             val tensorHandle: COpaquePointer = copy_from_blob_to_gpu_float(
                 array.toCValues(), shape.toCValues(), shape.size, device
             )!!
@@ -30,12 +37,24 @@ public abstract class TorchTensor<T, out TorchTensorBufferImpl : TorchTensorBuff
     }
 
     override fun toString(): String {
-        val nativeStringRepresentation: CPointer<ByteVar> = tensor_to_string(buffer.tensorHandle)!!
+        val nativeStringRepresentation: CPointer<ByteVar> = tensor_to_string(buffer.tensorHandle!!)!!
         val stringRepresentation = nativeStringRepresentation.toKString()
         dispose_char(nativeStringRepresentation)
         return stringRepresentation
     }
 
+    internal abstract fun wrap(
+        outStrides: TorchTensorStrides,
+        outScope: DeferScope,
+        outTensorHandle: COpaquePointer
+    ): TorchTensor<T, TorchTensorBufferImpl>
+
+    public fun copy(): TorchTensor<T, TorchTensorBufferImpl> = wrap(
+        outStrides = strides,
+        outScope = buffer.scope,
+        outTensorHandle = copy_tensor(buffer.tensorHandle!!)!!
+    )
+
 }
 
 public class TorchTensorFloat internal constructor(
@@ -44,6 +63,10 @@ public class TorchTensorFloat internal constructor(
     tensorHandle: COpaquePointer
 ) : TorchTensor<Float, TorchTensorBufferFloat>() {
     override val buffer: TorchTensorBufferFloat = TorchTensorBufferFloat(scope, tensorHandle)
+    override fun wrap(outStrides: TorchTensorStrides, outScope: DeferScope, outTensorHandle: COpaquePointer) =
+        TorchTensorFloat(
+            strides = outStrides, scope = outScope, tensorHandle = outTensorHandle
+        )
 }
 
 public class TorchTensorInt internal constructor(
@@ -52,6 +75,10 @@ public class TorchTensorInt internal constructor(
     tensorHandle: COpaquePointer
 ) : TorchTensor<Int, TorchTensorBufferInt>() {
     override val buffer: TorchTensorBufferInt = TorchTensorBufferInt(scope, tensorHandle)
+    override fun wrap(outStrides: TorchTensorStrides, outScope: DeferScope, outTensorHandle: COpaquePointer) =
+        TorchTensorInt(
+            strides = outStrides, scope = outScope, tensorHandle = outTensorHandle
+        )
 }
 
 public class TorchTensorFloatGPU internal constructor(
@@ -60,5 +87,9 @@ public class TorchTensorFloatGPU internal constructor(
     tensorHandle: COpaquePointer
 ) : TorchTensor<Float, TorchTensorBufferFloatGPU>() {
     override val buffer: TorchTensorBufferFloatGPU = TorchTensorBufferFloatGPU(scope, tensorHandle)
+    override fun wrap(outStrides: TorchTensorStrides, outScope: DeferScope, outTensorHandle: COpaquePointer) =
+        TorchTensorFloatGPU(
+            strides = outStrides, scope = outScope, tensorHandle = outTensorHandle
+        )
 }
 

kmath-torch/src/nativeMain/kotlin/kscience/kmath/torch/TorchTensorBuffer.kt

@@ -7,10 +7,13 @@ import ctorch.*
 
 public abstract class TorchTensorBuffer<T> internal constructor(
     internal val scope: DeferScope,
-    internal val tensorHandle: COpaquePointer
+    internal var tensorHandle: COpaquePointer?
 ) : MutableBuffer<T> {
 
-    override val size: Int = get_numel(tensorHandle)
+    override val size: Int
+        get(){
+            return get_numel(tensorHandle!!)
+        }
 
     init {
         scope.defer(::close)
@@ -18,13 +21,14 @@ public abstract class TorchTensorBuffer<T> internal constructor(
 
     protected fun close() {
         dispose_tensor(tensorHandle)
+        tensorHandle = null
     }
 
     internal abstract fun wrap(outScope: DeferScope, outTensorHandle: COpaquePointer): TorchTensorBuffer<T>
 
     override fun copy(): TorchTensorBuffer<T> = wrap(
         outScope = scope,
-        outTensorHandle = copy_tensor(tensorHandle)!!
+        outTensorHandle = copy_tensor(tensorHandle!!)!!
     )
 }
 
@@ -33,7 +37,10 @@ public class TorchTensorBufferFloat internal constructor(
     tensorHandle: COpaquePointer
 ) : TorchTensorBuffer<Float>(scope, tensorHandle) {
 
-    private val tensorData: CPointer<FloatVar> = get_data_float(tensorHandle)!!
+    private val tensorData: CPointer<FloatVar>
+        get(){
+            return get_data_float(tensorHandle!!)!!
+        }
 
     override operator fun get(index: Int): Float = tensorData[index]
 
@@ -55,7 +62,10 @@ public class TorchTensorBufferInt internal constructor(
     tensorHandle: COpaquePointer
 ) : TorchTensorBuffer<Int>(scope, tensorHandle) {
 
-    private val tensorData: CPointer<IntVar> = get_data_int(tensorHandle)!!
+    private val tensorData: CPointer<IntVar>
+        get(){
+            return get_data_int(tensorHandle!!)!!
+        }
 
     override operator fun get(index: Int): Int = tensorData[index]
 
@@ -76,14 +86,14 @@ public class TorchTensorBufferFloatGPU internal constructor(
     tensorHandle: COpaquePointer
 ) : TorchTensorBuffer<Float>(scope, tensorHandle) {
 
-    override operator fun get(index: Int): Float = get_at_offset_float(tensorHandle, index)
+    override operator fun get(index: Int): Float = get_at_offset_float(tensorHandle!!, index)
 
     override operator fun set(index: Int, value: Float) {
-        set_at_offset_float(tensorHandle, index, value)
+        set_at_offset_float(tensorHandle!!, index, value)
     }
 
     override operator fun iterator(): Iterator<Float> {
-        val cpuCopy = copy_to_cpu(tensorHandle)!!
+        val cpuCopy = copy_to_cpu(tensorHandle!!)!!
         val tensorCpuData = get_data_float(cpuCopy)!!
         val iteratorResult = (1..size).map { tensorCpuData[it - 1] }.iterator()
         dispose_tensor(cpuCopy)

kmath-torch/src/nativeTest/kotlin/kscience/kmath/torch/TestTorchTensor.kt

@@ -11,7 +11,7 @@ internal class TestTorchTensor {
     @Test
     fun intTensorLayout() = memScoped {
         val array = (1..24).toList().toIntArray()
-        val shape = intArrayOf(3, 2, 4)
+        val shape = intArrayOf(4, 6)
         val tensor = TorchTensor.copyFromIntArray(scope = this, array = array, shape = shape)
         tensor.elements().forEach {
             assertEquals(tensor[it.first], it.second)
@@ -30,4 +30,16 @@ internal class TestTorchTensor {
         assertTrue(tensor.buffer.contentEquals(array.asBuffer()))
     }
 
+    @Test
+    fun mutableStructure() = memScoped {
+        val array = (1..10).map { 1f * it }.toList().toFloatArray()
+        val shape = intArrayOf(10)
+        val tensor = TorchTensor.copyFromFloatArray(this, array, shape)
+        val tensorCopy = tensor.copy()
+
+        tensor[intArrayOf(0)] = 99f
+        assertEquals(99f, tensor[intArrayOf(0)])
+        assertEquals(1f, tensorCopy[intArrayOf(0)])
+    }
+
 }

kmath-torch/src/nativeTest/kotlin/kscience/kmath/torch/TestUtils.kt

@@ -1,9 +1,7 @@
 package kscience.kmath.torch
 
-import kotlin.test.Ignore
 import kotlin.test.Test
 import kotlin.test.assertEquals
-import kotlin.test.assertTrue
 
 
 internal class TestUtils {