# RUN: tf-mlir-translate -graphdef-to-mlir -tf-enable-shape-inference-on-import=false %s -tf-input-arrays=input0,input1 -tf-input-data-types=DT_INT32,DT_INT32 -tf-input-shapes=10:10 -tf-output-arrays=Sub -o - | FileCheck %s

node {
  name: "Add"
  op: "Add"
  input: "input0"
  input: "input1"
  # If device type or id doesn't exist, assign a default one (device:CPU:0).
  device: "/job:localhost/replica:0/task:0"
  attr {
    key: "T"
    value {

    key: "predict0"
    value: {
      inputs: {
        key: "inputs"
        value: {
          name: "input0"
          dtype: DT_STRING
        }
      }
      outputs: {
        key: "outputs"
        value: {
          name: "batch_func:0"
        }
      }
    }
  }
  signature_def: {
    key: "predict1"
    value: {
      inputs: {
        key: "tf_example_input"

// Test the output from TPU op is properly awaited before its use by map_fn.
// CHECK-LABEL: @main
// CHECK-SAME: ([[input0:%.*]]: !tf_mlrt.tensor, [[input1:%.*]]: !tf_mlrt.tensor)
func.func @main(%input0: tensor<i32>, %input1: tensor<i32>, %input2: tensor<!tf_type.variant<tensor<*xf32>>> ) -> tensor<i32> {
  %0 = "tf.Cast"(%input0) {__op_key = 0: i32, device = "/device:CPU:0"} : (tensor<i32>) -> tensor<f32>
  // CHECK: tf_mlrt_tpu.compile_and_execute

package hash_test

import (
	"bytes"
	"crypto/sha256"
	"encoding"
	"fmt"
	"log"
)

func Example_binaryMarshaler() {
	const (
		input1 = "The tunneling gopher digs downwards, "
		input2 = "unaware of what he will find."
	)

	first := sha256.New()
	first.Write([]byte(input1))

	marshaler, ok := first.(encoding.BinaryMarshaler)
	if !ok {
		log.Fatal("first does not implement encoding.BinaryMarshaler")
	}

// Tests input and output names from FlatBuffer are added to `tf.entry_function` attribute.

// CHECK-LABEL: @main
func.func @main(%arg0: tensor<4xi8>, %arg1: tensor<4xi32>) -> (tensor<4xi32>, tensor<4xi8>)
// CHECK: attributes {tf.entry_function = {inputs = "input0,input1", outputs = "output0,output1"}}
attributes {tf.entry_function = {inputs = "input0,input1", outputs = "output0,output1"}} {

// CHECK: {tf_saved_model.index_path = ["input2"]}
// CHECK-SAME: {tf_saved_model.index_path = ["input1"]}
// CHECK-SAME: {tf_saved_model.index_path = ["start_logits"]}
// CHECK-SAME: {tf_saved_model.index_path = ["end_logits"]}
// CHECK-SAME: tf.entry_function = {inputs = "serving_default_input2:0,serving_default_input1:0", outputs = "StatefulPartitionedCall:1,StatefulPartitionedCall:0"}

    RunOptions run_options;

    Tensor input_a = CreateInputTensor(shape, 0);
    Tensor input_b = CreateInputTensor(shape, shape.num_elements());
    Tensor input_c = CreateInputTensor(shape, 2 * shape.num_elements());
    TF_RETURN_IF_ERROR(session->Run(
        run_options,
        {std::make_pair("a", input_a), std::make_pair("b", input_b),
         std::make_pair("c", input_c)},
        {"m"}, {}, &golden_output_tensors, nullptr));

// CHECK-LABEL: @main_stream_0
// CHECK-SAME: ([[input0:%.*]]: !tf_mlrt.tensor, [[promise_b:%.*]]: !mlrt.promise)
func.func @main_stream_0(%input0: tensor<i32>, %promise_b: !mlrt.promise) {
  %const = "tf.Const"() {__op_key = 0 : i32, value = dense<1> : tensor<i32>} : () -> tensor<i32>
  // CHECK: [[a:%.*]] = tf_mlrt.executeop([[input0]],
  // CHECK-SAME: AddV2

    def "can read from stream and then restart to use another stream"() {
        def input1 = encoded("string 1")
        def input2 = encoded("string 2")

        given:
        def decoder = new KryoBackedDecoder(input1)
        decoder.readString()
        decoder.restart(input2)

        expect:
        decoder.readPosition == 0
        decoder.readString() == "string 2"

  Init(2);
  // Create nested loop:
  //  while (input1 < 6) {
  //    inner_input1 = input1
  //    while (inner_input1 < 3) {
  //      input2 += 1
  //      inner_input1 += 2
  //    }
  //    input1 += input2
  //  }
  //
  // Expected execution with initial values input1 = input2 = 0:
  //
  // outer inner               inner_
  // step# step# input1 input2 input1
  // ------------------------------------