tf.data abstraction in IDatasetV2.

src/TensorFlowNET.Core/Data/BatchDataset.cs

@@ -28,7 +28,7 @@ namespace Tensorflow
             }
             else
             {
-                _structure = input_dataset.element_spec.Select(x => x._batch(-1)).ToArray();
+                structure = input_dataset.element_spec.Select(x => x._batch(-1)).ToArray();
             }
 
             variant_tensor = ops.batch_dataset_v2(input_dataset.variant_tensor,

src/TensorFlowNET.Core/Data/DatasetV2.cs

@@ -15,13 +15,13 @@ namespace Tensorflow
         protected dataset_ops ops = new dataset_ops();
         public Tensor variant_tensor { get; set; }
 
-        public TensorSpec[] _structure { get; set; }
+        public TensorSpec[] structure { get; set; }
 
-        public TensorShape[] output_shapes => _structure.Select(x => x.shape).ToArray();
+        public TensorShape[] output_shapes => structure.Select(x => x.shape).ToArray();
         
-        public TF_DataType[] output_types => _structure.Select(x => x.dtype).ToArray();
+        public TF_DataType[] output_types => structure.Select(x => x.dtype).ToArray();
         
-        public TensorSpec[] element_spec => _structure;
+        public TensorSpec[] element_spec => structure;
 
         public IDatasetV2 take(int count = -1)
             => new TakeDataset(this, count: count);
@@ -37,13 +37,52 @@ namespace Tensorflow
 
         public IDatasetV2 shuffle(int buffer_size, int? seed = null, bool reshuffle_each_iteration = true)
             => new ShuffleDataset(this, buffer_size, seed: seed, reshuffle_each_iteration: reshuffle_each_iteration);
-        
+
+        public IDatasetV2 optimize(string[] optimizations, string[] optimization_configs)
+            => new OptimizeDataset(this, optimizations, optimization_configs: optimization_configs);
+
+        public IDatasetV2 model(AutotuneAlgorithm algorithm, long cpu_budget)
+            => new ModelDataset(this, algorithm, cpu_budget);
+
+        public IDatasetV2 apply_options()
+        {
+            // (1) Apply threading options
+            var graph_rewrites = new[]
+            {
+                "map_and_batch_fusion",
+                "noop_elimination",
+                "shuffle_and_repeat_fusion"
+            };
+
+            var graph_rewrite_configs = new string[0];
+
+            // (2) Apply graph rewrite options
+            var dataset = optimize(graph_rewrites, graph_rewrite_configs);
+
+            // (3) Apply autotune options
+            var autotune = true;
+            long cpu_budget = 0;
+
+            if (autotune)
+                dataset = dataset.model(AutotuneAlgorithm.HILL_CLIMB, cpu_budget);
+
+            // (4) Apply stats aggregator options
+
+            return dataset;
+        }
+
         public override string ToString()
-            => $"{GetType().Name} shapes: ({_structure[0].shape}, {_structure[1].shape}), types: (tf.{_structure[0].dtype.as_numpy_name()}, tf.{_structure[1].dtype.as_numpy_name()})";
+            => $"{GetType().Name} shapes: ({structure[0].shape}, {structure[1].shape}), types: (tf.{structure[0].dtype.as_numpy_name()}, tf.{structure[1].dtype.as_numpy_name()})";
 
         public IEnumerator<(Tensor, Tensor)> GetEnumerator()
         {
-            throw new NotImplementedException();
+            var ownedIterator = new OwnedIterator(this);
+
+            Tensor[] results = ownedIterator.next();
+            while (results != null)
+            {
+                yield return (results[0], results[1]);
+            }
         }
 
         IEnumerator IEnumerable.GetEnumerator()

src/TensorFlowNET.Core/Data/IDatasetV2.cs

@@ -15,7 +15,7 @@ namespace Tensorflow
 
         TensorSpec[] element_spec { get; }
 
-        TensorSpec[] _structure { get; set; }
+        TensorSpec[] structure { get; set; }
 
         /// <summary>
         /// 
@@ -31,5 +31,15 @@ namespace Tensorflow
         IDatasetV2 prefetch(int buffer_size = -1, int? slack_period = null);
 
         IDatasetV2 take(int count);
+
+        IDatasetV2 optimize(string[] optimizations, string[] optimization_configs);
+
+        IDatasetV2 model(AutotuneAlgorithm algorithm, long cpu_budget);
+
+        /// <summary>
+        /// Apply options, such as optimization configuration, to the dataset.
+        /// </summary>
+        /// <returns></returns>
+        IDatasetV2 apply_options();
     }
 }

src/TensorFlowNET.Core/Data/IteratorBase.cs

@@ -0,0 +1,10 @@
+using System;
+using System.Collections.Generic;
+using System.Text;
+
+namespace Tensorflow
+{
+    public class IteratorBase
+    {
+    }
+}

src/TensorFlowNET.Core/Data/IteratorResourceDeleter.cs

@@ -0,0 +1,29 @@
+using System;
+using System.Collections.Generic;
+using System.Text;
+using static Tensorflow.Binding;
+
+namespace Tensorflow
+{
+    /// <summary>
+    /// An object which cleans up an iterator resource handle.
+    /// </summary>
+    public class IteratorResourceDeleter : IDisposable
+    {
+        Tensor _handle;
+        Tensor _deleter;
+        dataset_ops ops;
+
+        public IteratorResourceDeleter(Tensor handle, Tensor deleter)
+        {
+            _handle = handle;
+            _deleter = deleter;
+            ops = new dataset_ops();
+        }
+
+        public void Dispose()
+        {
+            ops.delete_iterator(_handle, _deleter);
+        }
+    }
+}

src/TensorFlowNET.Core/Data/ModelDataset.cs

@@ -0,0 +1,25 @@
+using System;
+using System.Collections.Generic;
+using System.Text;
+using Tensorflow.Framework.Models;
+
+namespace Tensorflow
+{
+    /// <summary>
+    /// A `Dataset` that acts as an identity, and models performance.
+    /// </summary>
+    public class ModelDataset : UnaryUnchangedStructureDataset
+    {
+        public ModelDataset(IDatasetV2 input_dataset, 
+            AutotuneAlgorithm algorithm, 
+            long cpu_budget) :
+            base(input_dataset)
+        {
+            variant_tensor = ops.model_dataset(input_dataset.variant_tensor,
+                output_types,
+                output_shapes,
+                algorithm,
+                cpu_budget);
+        }
+    }
+}

src/TensorFlowNET.Core/Data/OptimizeDataset.cs

@@ -0,0 +1,34 @@
+using System;
+using System.Collections.Generic;
+using System.Text;
+using static Tensorflow.Binding;
+
+namespace Tensorflow
+{
+    /// <summary>
+    /// A `Dataset` that acts as an identity, and applies optimizations.
+    /// </summary>
+    public class OptimizeDataset : UnaryUnchangedStructureDataset
+    {
+        Tensor _optimizations;
+
+        public OptimizeDataset(IDatasetV2 dataset, 
+            string[] optimizations = null, 
+            string[] optimization_configs = null) : 
+            base(dataset)
+        {
+            if (optimizations == null)
+                optimizations = new string[0];
+            if (optimization_configs == null)
+                optimization_configs = new string[0];
+
+            _optimizations = tf.convert_to_tensor(optimizations, dtype: TF_DataType.TF_STRING, name: "optimizations");
+            variant_tensor = ops.optimize_dataset(
+                _input_dataset.variant_tensor,
+                _optimizations,
+                output_types, 
+                output_shapes,
+                optimization_configs: optimization_configs);
+        }
+    }
+}

src/TensorFlowNET.Core/Data/OwnedIterator.cs

@@ -0,0 +1,40 @@
+using System;
+using System.Collections.Generic;
+using System.Text;
+using Tensorflow.Framework.Models;
+
+namespace Tensorflow
+{
+    /// <summary>
+    /// An iterator producing tf.Tensor objects from a tf.data.Dataset.
+    /// </summary>
+    public class OwnedIterator : IteratorBase
+    {
+        IDatasetV2 _dataset;
+        TensorSpec[] _element_spec;
+        dataset_ops ops = new dataset_ops();
+        Tensor _iterator_resource;
+        Tensor _deleter;
+        IteratorResourceDeleter _resource_deleter;
+
+        public OwnedIterator(IDatasetV2 dataset)
+        {
+            _create_iterator(dataset);
+        }
+
+        void _create_iterator(IDatasetV2 dataset)
+        {
+            dataset = dataset.apply_options();
+            _dataset = dataset;
+            _element_spec = dataset.element_spec;
+            (_iterator_resource, _deleter) = ops.anonymous_iterator_v2(_dataset.output_types, _dataset.output_shapes);
+            ops.make_iterator(dataset.variant_tensor, _iterator_resource);
+
+            // Delete the resource when this object is deleted
+            _resource_deleter = new IteratorResourceDeleter(_iterator_resource, _deleter);
+        }
+
+        public Tensor[] next()
+            => ops.iterator_get_next(_iterator_resource, _dataset.output_types, _dataset.output_shapes);
+    }
+}

src/TensorFlowNET.Core/Data/TensorSliceDataset.cs

@@ -15,7 +15,7 @@ namespace Tensorflow
         {
             _tensors = new[] { tf.convert_to_tensor(features), tf.convert_to_tensor(labels) };
             var batched_spec = _tensors.Select(x => x.ToTensorSpec()).ToArray();
-            _structure = batched_spec.Select(x => x._unbatch()).ToArray();
+            structure = batched_spec.Select(x => x._unbatch()).ToArray();
             
             variant_tensor = ops.tensor_slice_dataset(_tensors, output_shapes);
         }

src/TensorFlowNET.Core/Data/UnaryDataset.cs

@@ -15,7 +15,7 @@ namespace Tensorflow
         public UnaryDataset(IDatasetV2 input_dataset)
         {
             _input_dataset = input_dataset;
-            _structure = input_dataset._structure;
+            structure = input_dataset.structure;
         }
     }
 }