// RUN: stablehlo-quant-opt %s -split-input-file -stablehlo-unfuse-mhlo-batch-norm | FileCheck %s

// CHECK-LABEL: @unfuse_batch_norm
// CHECK-SAME: %[[X:[^:[:space:]]+]]
// CHECK-SAME: %[[SCALE:[^:[:space:]]+]]
// CHECK-SAME: %[[OFFSET:[^:[:space:]]+]]
// CHECK-SAME: %[[MEAN:[^:[:space:]]+]]
// CHECK-SAME: %[[VARIANCE:[^:[:space:]]+]]
func.func @unfuse_batch_norm(
    %x: tensor<4x256xf32>, %scale: tensor<256xf32>, %offset: tensor<256xf32>,

// CHECK-DAG: %[[CONST:.*]] = "tf.Const"() <{value = dense<0.000000e+00> : tensor<512x512xf32>}> : () -> tensor<512x512xf32>
// CHECK: %[[PARTITIONEDCALL:.*]] = "tf.PartitionedCall"(%arg0, %[[CONST]])
// CHECK-SAME: <{config = "",
// CHECK-SAME: f = @composite_matmul_fn_1}>
// CHECK-NOT: {_tfl_quant_trait = "fully_quantizable"
// CHECK: %[[UNQUANTIZED_OUTPUT:.*]] = "tf.MatMul"(%arg0, %arg0)
// CHECK: }

// CHECK-LABEL: private @composite_matmul_fn_1

// CHECK-SAME: %[[ARG_0:.*]]: tensor<!tf_type.variant>, %[[ARG_1:.*]]: tensor<i64>
func.func @single_state_single_dataset_type_no_arguments(
      %arg0: tensor<!tf_type.variant>,
      %arg1: tensor<i64>
    ) {
    // CHECK:      %[[ANON_ITER:[0-9]*]] = "tf.AnonymousIteratorV3"
    // CHECK-SAME: output_shapes = [#tf_type.shape<32>]
    // CHECK-SAME: output_types = [f32]

        id.getDisplayName() == "original -> current (foo)"
    }

    def "equals and hash code differentiate between same and different instances"() {
        def componentId = newComponentId("foo")

        when:
        def id = new TransformedComponentFileArtifactIdentifier(componentId, "a", "b")
        def same = new TransformedComponentFileArtifactIdentifier(componentId, "a", "b")

    }

    def "repositories have the same id when metadata sources and other configuration is the same"() {
        def repo = newRepo()
        def same = newRepo()
        def different = newRepo()

        given:
        repo.url = new URI("http://localhost")
        source(repo)
        same.url = new URI("http://localhost")
        source(same)
        different.url = new URI("http://localhost")

//CHECK: %0 = "vhlo.convolution_v1"(%arg0, %arg1) <{
//CHECK-SAME: batch_group_count = #vhlo.integer_v1<1 : i64>,
//CHECK-SAME: feature_group_count = #vhlo.integer_v1<32 : i64>,
//CHECK-SAME: input_batch_dimension = #vhlo.integer_v1<0 : i64>,
//CHECK-SAME: input_feature_dimension = #vhlo.integer_v1<3 : i64>,
//CHECK-SAME: input_spatial_dimensions = #vhlo.tensor_v1<dense<[1, 2]> : tensor<2xi64>>,
//CHECK-SAME: kernel_input_feature_dimension = #vhlo.integer_v1<2 : i64>,

Did you know?

  - [testing.T] is one doc link.
  - So is [testing.M].
  - So is [testing.B].
    This is the same list paragraph.

    There is [testing.PB] in this list item, too!
-- text --
Did you know?

  - testing.T is one doc link.

  - So is testing.M.

  - So is testing.B. This is the same list paragraph.

    There is testing.PB in this list item, too!
-- markdown --
Did you know?

# Tests different SignatureDef's with identical method_name string

# CHECK:      func {{@[a-zA-Z_0-9]+}}(
# CHECK-SAME:   {{.*}})
# CHECK-SAME: attributes {{.*}} tf_saved_model.exported_names = ["key"]

# CHECK:      func {{@[a-zA-Z_0-9]+}}(
# CHECK-SAME:   {{.*}})
# CHECK-SAME: attributes {{.*}} tf_saved_model.exported_names = ["key2"]


def Test():

  x = tf.constant(1.0, shape=(3, 3))

// CHECK-LABEL: func private @main_stream_{{[0-9]*}}
// CHECK-SAME: ([[ARG:%.*]]: tensor<i32>, [[FUTURE:%.*]]: !mlrt.future, [[CONTROL_PROMISE:%.*]]: !mlrt.promise)
// CHECK: [[HANDLE:%.*]] = "tf_mlrt.tf_await"([[FUTURE]])
// CHECK: "tf.AssignVariableOp"([[HANDLE]], [[ARG]])
// CHECK-NEXT: mlrt.promise_control [[CONTROL_PROMISE]]

// CHECK-LABEL: func private @main_stream_{{[0-9]*}}

[[concepts_repeatable_task_outputs]]
== Repeatable task outputs

Assume that you have a code generator task as part of your build.
When you have a fully up to date build and you clean and re-run the code generator task on the same code base it should generate _exactly the same output_, so anything that depends on that output will stay up-to-date.