[Type Refactor] Merge TType and Tensor instances as a single entity (#160)

* Merge TType and Tensor instances as a single entity

* Rectify documentation based on PR review

* Rebase on master

Author: karllessard

tensorflow-core/tensorflow-core-api/src/bazel/op_generator/op_specs.cc

@@ -152,8 +152,7 @@ std::pair<Type, Type> TypeResolver::TypesOf(const OpDef_AttrDef& attr_def,
     types = MakeTypePair(Type::Class("Shape", "org.tensorflow.ndarray"));
 
   } else if (attr_type == "tensor") {
-    types = MakeTypePair(Type::Class("Tensor", "org.tensorflow")
-                             .add_parameter(Type::Wildcard()));
+    types = MakeTypePair(Type::Class("Tensor", "org.tensorflow"));
 
   } else if (attr_type == "type") {
     Type type = *iterable_out ? Type::Wildcard() : NextGeneric();

tensorflow-core/tensorflow-core-api/src/gen/annotations/org/tensorflow/op/DebuggingOps.java

@@ -24,7 +24,6 @@
 import org.tensorflow.EagerSession;
 import org.tensorflow.ExecutionEnvironment;
 import org.tensorflow.Operand;
-import org.tensorflow.Tensor;
 import org.tensorflow.ndarray.BooleanNdArray;
 import org.tensorflow.ndarray.ByteNdArray;
 import org.tensorflow.ndarray.DoubleNdArray;
@@ -349,10 +348,10 @@ public final class Ops {
 
   public final SignalOps signal;
 
-  public final TrainOps train;
-
   public final QuantizationOps quantization;
 
+  public final TrainOps train;
+
   private final Scope scope;
 
   private Ops(Scope scope) {
@@ -374,8 +373,8 @@ private Ops(Scope scope) {
     math = new MathOps(this);
     audio = new AudioOps(this);
     signal = new SignalOps(this);
-    train = new TrainOps(this);
     quantization = new QuantizationOps(this);
+    train = new TrainOps(this);
   }
 
   /**
@@ -1708,17 +1707,6 @@ public Constant<TInt64> constant(Shape shape) {
     return Constant.tensorOf(scope, shape);
   }
 
-  /**
-   * Create a constant from a Tensor.
-   *
-   * @param scope is a scope used to add the underlying operation.
-   * @param tensor a Tensor holding the constant value
-   * @return a constant of the same data type as `tensor`
-   */
-  public <T extends TType> Constant<T> constant(Tensor<T> tensor) {
-    return Constant.create(scope, tensor);
-  }
-
   /**
    * Creates a constant of {@code String} elements, using the given charset.
    *
@@ -1879,6 +1867,20 @@ public <T extends TType> Constant<T> constant(DataType<T> type, Shape shape,
     return Constant.tensorOf(scope, type, shape, data);
   }
 
+  /**
+   * Create a constant by making an immutable copy of {@code tensor}.
+   *
+   *  <p>Note: this endpoint cannot be simply called {@code constant} since it will conflict with
+   *  other endpoints accepting an NdArray in parameter {e.g. {@link #tensorOf(Scope, FloatNdArray)}}.
+   *
+   * @param scope is a scope used to add the underlying operation.
+   * @param tensor a Tensor holding the constant value
+   * @return a constant of the same data type as `tensor`
+   */
+  public <T extends TType> Constant<T> constantOf(T tensor) {
+    return Constant.create(scope, tensor);
+  }
+
   /**
    * This op consumes a lock created by `MutexLock`.
    *  <p>

tensorflow-core/tensorflow-core-api/src/gen/annotations/org/tensorflow/op/Ops.java

@@ -99,13 +99,13 @@ public Options maxImages(Long maxImages) {
     /**
      * @param badColor Color to use for pixels with non-finite values.
      */
-    public Options badColor(Tensor<?> badColor) {
+    public Options badColor(Tensor badColor) {
       this.badColor = badColor;
       return this;
     }
 
     private Long maxImages;
-    private Tensor<?> badColor;
+    private Tensor badColor;
 
     private Options() {
     }
@@ -150,7 +150,7 @@ public static Options maxImages(Long maxImages) {
   /**
    * @param badColor Color to use for pixels with non-finite values.
    */
-  public static Options badColor(Tensor<?> badColor) {
+  public static Options badColor(Tensor badColor) {
     return new Options().badColor(badColor);
   }
 

tensorflow-core/tensorflow-core-api/src/gen/java/org/tensorflow/op/summary/ImageSummary.java

@@ -86,5 +86,5 @@ public String toString() {
    * @param outputIdx index of the output of this operation
    * @return output tensor
    */
-  abstract Tensor<?> tensor(int outputIdx);
+  abstract Tensor tensor(int outputIdx);
 }

tensorflow-core/tensorflow-core-api/src/main/java/org/tensorflow/AbstractOperation.java

@@ -34,7 +34,7 @@
  *
  * <pre>{@code
  * ConcreteFunction myFunction = savedModelBundle.function("myFunctionSignatureName");
- * Map<String, Tensor<?>> outputTensorMap = myFunction.call(inputTensorMap);
+ * Map<String, Tensor> outputTensorMap = myFunction.call(inputTensorMap);
  * }</pre>
  */
 public class ConcreteFunction implements AutoCloseable {
@@ -61,8 +61,8 @@ public class ConcreteFunction implements AutoCloseable {
    *
    *   public static void main(String args[]) {
    *     try (ConcreteFunction function = ConcreteFunction.create(MyModel::addTwo);
-   *         Tensor<TFloat32> x = TFloat32.scalarOf(2.0f)) {
-   *       assertEquals(4.0f, function.call(x).expect(TFloat32.DTYPE).data().getFloat());
+   *         TFloat32 x = TFloat32.scalarOf(2.0f)) {
+   *       assertEquals(4.0f, ((TFloat32)function.call(x)).getFloat());
    *     }
    *   }
    * }
@@ -97,8 +97,8 @@ public static ConcreteFunction create(Function<Ops, Signature> functionBuilder)
    *   Signature signature = Signature.builder().input("x", input).output("y", output).build();
    *
    *   try (ConcreteFunction f = ConcreteFunction.create(signature, g);
-   *       Tensor<TFloat32> x = TFloat32.scalarOf(2.0f)) {
-   *     assertEquals(4.0f, function.call(x).expect(TFloat32.DTYPE).data().getFloat());
+   *       TFloat32 x = TFloat32.scalarOf(2.0f)) {
+   *     assertEquals(4.0f, ((TFloat32)function.call(x)).getFloat());
    *   }
    *   // Graph g is still valid at this point
    * }
@@ -129,8 +129,8 @@ public static ConcreteFunction create(Signature signature, Graph graph) {
    *     // Auto-closing the function just as an example but this is not required since it has
    *     // no effect
    *     try (ConcreteFunction f = ConcreteFunction.create(signature, s);
-   *         Tensor<TFloat32> t = TFloat32.scalarOf(2.0f)) {
-   *       assertEquals(4.0f, function.call(x).expect(TFloat32.DTYPE).data().getFloat());
+   *         TFloat32 t = TFloat32.scalarOf(2.0f)) {
+   *       assertEquals(4.0f, ((TFloat32)function.call(x)).getFloat());
    *     }
    *     // Session s is still valid at this point
    *   }
@@ -163,14 +163,14 @@ public Signature signature() {
    * @return output tensors resulting from the execution of the function,
    *         mapped by their signature name
    */
-  public Map<String, Tensor<?>> call(Map<String, Tensor<?>> arguments)
+  public Map<String, Tensor> call(Map<String, Tensor> arguments)
       throws IllegalArgumentException {
 
     final SignatureDef signatureDef = signature.asSignatureDef();
     final Session.Runner runner = session.runner();
 
     signatureDef.getInputsMap().forEach((argName, t) -> {
-      Tensor<?> tensor = arguments.get(argName);
+      Tensor tensor = arguments.get(argName);
       if (tensor == null) {
         throw new IllegalArgumentException(String.format("Missing argument [%s]", argName));
       }
@@ -180,10 +180,10 @@ public Map<String, Tensor<?>> call(Map<String, Tensor<?>> arguments)
     Map<String, TensorInfo> outputToNode = signatureDef.getOutputsMap();
     outputToNode.values().forEach(t -> runner.fetch(t.getName()));
 
-    List<Tensor<?>> resultTensors = runner.run();
+    List<Tensor> resultTensors = runner.run();
     try {
-      ListIterator<Tensor<?>> resultTensorIter = resultTensors.listIterator();
-      Map<String, Tensor<?>> returnMap = new HashMap<String, Tensor<?>>();
+      ListIterator<Tensor> resultTensorIter = resultTensors.listIterator();
+      Map<String, Tensor> returnMap = new HashMap<String, Tensor>();
 
       // Use the output names as present in the signature definition
       for (String nodeName: outputToNode.keySet()) {
@@ -193,7 +193,7 @@ public Map<String, Tensor<?>> call(Map<String, Tensor<?>> arguments)
 
     } catch (Exception e) {
       // Release tensors before throwing exception
-      for (Tensor<?> t : resultTensors) {
+      for (Tensor t : resultTensors) {
         t.close();
       }
       throw e;
@@ -210,7 +210,7 @@ public Map<String, Tensor<?>> call(Map<String, Tensor<?>> arguments)
    * @throws IllegalArgumentException if there are multiple input or output parameters defined
    *                                  in the function
    */
-  public Tensor<?> call(Tensor<?> tensor) throws IllegalArgumentException {
+  public Tensor call(Tensor tensor) throws IllegalArgumentException {
     final SignatureDef signatureDef = signature.asSignatureDef();
 
     if (signatureDef.getInputsCount() != 1) {

tensorflow-core/tensorflow-core-api/src/main/java/org/tensorflow/ConcreteFunction.java

@@ -16,7 +16,6 @@
 package org.tensorflow;
 
 import org.tensorflow.internal.c_api.TF_Tensor;
-import org.tensorflow.ndarray.Shape;
 import org.tensorflow.types.TBfloat16;
 import org.tensorflow.types.TBool;
 import org.tensorflow.types.TFloat16;
@@ -35,13 +34,17 @@ public final class DataType<T extends TType> {
   public interface TensorMapper<T> {
 
     /**
-     * Maps tensor memory to a data structure for manipulating elements of this type.
+     * Maps the tensor memory to a n-dimensional typed data space.
      *
-     * @param nativeTensor pointer to the native tensor
-     * @param shape the shape of the tensor
-     * @return data structure of elements of this type
+     * <p>This method is designed to be invoked internally by this library only, in order to pass the
+     * native handle of {@code tensor} as {@code nativeHandle} (and since only classes from the
+     * {@code org.tensorflow} package can retrieve such handle).
+     *
+     * @param tensor the tensor to map in its raw nature
+     * @param nativeHandle native handle of the tensor
+     * @return a typed tensor of type {@code T}
      */
-    T apply(TF_Tensor nativeTensor, Shape shape);
+    T apply(RawTensor tensor, TF_Tensor nativeHandle);
   }
 
   /**
@@ -158,13 +161,13 @@ int nativeCode() {
   }
 
   /**
-   * Maps a tensor to a data structure for manipulating elements of this type.
+   * Maps a raw tensor to a typed tensor.
    *
    * @param tensor tensor to map
    * @return data structure of elements of this type
    */
-  T map(Tensor<T> tensor) {
-    return tensorMapper.apply(tensor.nativeHandle(), tensor.shape());
+  T map(RawTensor tensor) {
+    return tensorMapper.apply(tensor, tensor.nativeHandle());
   }
 
   private final int nativeCode;

tensorflow-core/tensorflow-core-api/src/main/java/org/tensorflow/DataType.java

@@ -91,7 +91,7 @@ public TFE_TensorHandle getUnsafeNativeHandle(int outputIndex) {
   public Shape shape(int outputIndex) {
     // If the tensor of this output has already been resolved, return its shape.
     // Otherwise, retrieve the tensor shape from the native library.
-    Tensor<?> tensor = outputTensors.get(outputIndex);
+    Tensor tensor = outputTensors.get(outputIndex);
     if (tensor != null) {
       return tensor.shape();
     }
@@ -107,7 +107,7 @@ public Shape shape(int outputIndex) {
   public DataType<?> dtype(int outputIndex) {
     // If the tensor of this output has already been resolved, return its datatype.
     // Otherwise, retrieve the tensor datatype from the native library.
-    Tensor<?> tensor = outputTensors.get(outputIndex);
+    Tensor tensor = outputTensors.get(outputIndex);
     if (tensor != null) {
       return tensor.dataType();
     }
@@ -116,8 +116,8 @@ public DataType<?> dtype(int outputIndex) {
   }
 
   @Override
-  public Tensor<?> tensor(int outputIndex) {
-    Tensor<?> tensor = outputTensors.get(outputIndex);
+  public Tensor tensor(int outputIndex) {
+    Tensor tensor = outputTensors.get(outputIndex);
     if (tensor == null) {
       tensor = resolveTensor(outputIndex);
     }
@@ -127,21 +127,21 @@ public Tensor<?> tensor(int outputIndex) {
   private final EagerSession session;
   private final String type;
   private final String name;
-  private final AtomicReferenceArray<Tensor<?>> outputTensors;
+  private final AtomicReferenceArray<Tensor> outputTensors;
 
-  private Tensor<?> resolveTensor(int outputIndex) {
+  private Tensor resolveTensor(int outputIndex) {
     // Take an optimistic approach, where we attempt to resolve the output tensor without locking.
     // If another thread has resolved it meanwhile, release our copy and reuse the existing one
     // instead.
-    Tensor<?> tensor = resolveTensorHandle(getUnsafeNativeHandle(outputIndex), session);
+    Tensor tensor = resolveTensorHandle(getUnsafeNativeHandle(outputIndex), session);
     if (!outputTensors.compareAndSet(outputIndex, null, tensor)) {
-      session.detach(tensor.nativeHandle());
+      session.detach(tensor.asRawTensor().nativeHandle());
       tensor = outputTensors.get(outputIndex);
     }
     return tensor;
   }
 
-  private TFE_Op opHandle;
+  private final TFE_Op opHandle;
   private final TFE_TensorHandle[] outputHandles;
 
   private static void requireOp(TFE_Op handle) {
@@ -156,13 +156,13 @@ private static void requireTensorHandle(TFE_TensorHandle handle) {
     }
   }
 
-  private static Tensor<?> resolveTensorHandle(TFE_TensorHandle handle, EagerSession session) {
+  private static Tensor resolveTensorHandle(TFE_TensorHandle handle, EagerSession session) {
     requireTensorHandle(handle);
     try (PointerScope scope = new PointerScope()) {
       TF_Status status = TF_Status.newStatus();
       TF_Tensor tensor = TFE_TensorHandleResolve(handle, status).withDeallocator();
       status.throwExceptionIfNotOK();
-      return Tensor.fromHandle(tensor, session);
+      return RawTensor.fromHandle(tensor, session).asTypedTensor();
     }
   }
 

tensorflow-core/tensorflow-core-api/src/main/java/org/tensorflow/EagerOperation.java

@@ -175,13 +175,13 @@ public EagerOperationBuilder setAttr(String name, DataType<?>[] values) {
   }
 
   @Override
-  public EagerOperationBuilder setAttr(String name, Tensor<?> value) {
-    setAttrTensor(opHandle, name, value.nativeHandle());
+  public EagerOperationBuilder setAttr(String name, Tensor value) {
+    setAttrTensor(opHandle, name, value.asRawTensor().nativeHandle());
     return this;
   }
 
   @Override
-  public EagerOperationBuilder setAttr(String name, Tensor<?>[] values) {
+  public EagerOperationBuilder setAttr(String name, Tensor[] values) {
     // TODO (karllessard) could be supported by adding this attribute type in the eager C API
     throw new UnsupportedOperationException(
         "Tensor list attributes are not supported in eager mode");

tensorflow-core/tensorflow-core-api/src/main/java/org/tensorflow/EagerOperationBuilder.java

@@ -158,7 +158,7 @@ DataType<?> dtype(int outputIdx) {
   }
 
   @Override
-  Tensor<?> tensor(int outputIdx) {
+  Tensor tensor(int outputIdx) {
     throw new IllegalStateException("Graph tensors must be fetched by running a session");
   }