if needHash {
						sum, err := dirhash.HashZip(zipfile, dirhash.Hash1)
						if err != nil {
							t.Errorf("repo.Zip(%q): %v", tt.version, err)
						} else if sum != tt.zipSum {
							t.Errorf("repo.Zip(%q): got file with sum %q, want %q", tt.version, sum, tt.zipSum)
						} else if zipFileHash := hex.EncodeToString(h.Sum(nil)); zipFileHash != tt.zipFileHash {

)

const zipSumsPath = "testdata/zip_sums.csv"

type zipSumTest struct {
	m                     module.Version
	wantSum, wantFileHash string
}

func TestZipSums(t *testing.T) {
	if !*enableZipSum {
		// This test is very slow and heavily dependent on external repositories.
		// Only run it explicitly.
		t.Skip("TestZipSum not enabled with -zipsum")
	}
	if *shardCount < 1 {

import org.gradle.integtests.fixtures.SourceFile

import static org.gradle.nativeplatform.fixtures.app.SourceFileElement.ofFile

class CppSum extends CppSourceFileElement implements SumElement {
    final SourceFileElement header
    final SourceFileElement source

    CppSum(String publicHeadersDir = "headers") {
        header = ofFile(new SourceFile(publicHeadersDir, "sum.h", """
#ifdef _WIN32

// RUN: tf-opt -split-input-file -verify-diagnostics -tf-einsum %s | FileCheck %s

func.func @unary_einsum_reduce_sum_transpose(%arg0: tensor<3x4x5x6xf32>) -> tensor<3x5x4xf32> {
  %0 = "tf.Einsum"(%arg0) {T = "tfdtype$DT_FLOAT", equation = "...gse->...sg"}: (tensor<3x4x5x6xf32>) -> tensor<3x5x4xf32>
  func.return %0 : tensor<3x5x4xf32>
  // CHECK-LABEL: unary_einsum_reduce_sum_transpose
  // CHECK-DAG: %[[cst:.*]] = arith.constant dense<3> : tensor<1xi32>

}

struct EinsumDimensionNumbers {
  // Each field contains the list of dimensions appearing only in the specifed
  // arguments of the einsum op with natural ordering. For example `rhs_out`
  // contains the dimensions appearing in the RHS and the OUTPUT of the einsum
  // but not in the LHS.
  std::vector<int64_t> lhs;
  std::vector<int64_t> rhs;
  std::vector<std::tuple<int64_t, int64_t>> lhs_rhs;

  auto z = Einsum(scope_, {x}, "ii->");
  TensorShape z_shape({});
  RunTest({x}, {x_shape}, {z}, {z_shape});
}

TEST_F(LinalgGradTest, Einsum_TraceBroadcast) {
  TensorShape x_shape({4, 3, 3});
  Output x = Placeholder(scope_, DT_FLOAT, Placeholder::Shape(x_shape));
  // Note: In Python this could just be "ii" becuase tf.einsum normalizes the
  // equation, but c++ doesn't do that.
  auto z = Einsum(scope_, {x}, "...ii->...");

// RUN: tf-opt %s -tf-batch-matmul-to-tf-einsum | FileCheck %s

func.func @test_batch_matmul_to_einsum(%arg0: tensor<1x2x3xf32>, %arg1: tensor<3x4xf32>) -> tensor<1x2x4xf32> {
  // CHECK-LABEL: test_batch_matmul_to_einsum
  // CHECK: "tf.Einsum"(%arg0, %arg1) <{equation = "...mk,...kn->...mn"}> : (tensor<1x2x3xf32>, tensor<3x4xf32>) -> tensor<1x2x4xf32>

### Format

The format for ADR should follow this template:

```markdown
# ADR-000X - Title

## Date

20YY-MM-DD

## Context

Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna
aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat.

public class SourceSinkTester<S, T, F extends SourceSinkFactory<S, T>> extends TestCase {

  static final String LOREM_IPSUM =
      "Lorem ipsum dolor sit amet, consectetur adipiscing "
          + "elit. Cras fringilla elit ac ipsum adipiscing vulputate. Maecenas in lorem nulla, ac "
          + "sollicitudin quam. Praesent neque elit, sodales quis vestibulum vel, pellentesque nec "

  // Claim: For the einsum operation z = einsum("{eq_x},{eq_y}->{eq_z}", x, y),
  //   where the equation involves only Tensor contractions, generalized traces
  //   and transposes, the input gradients are given by the vector-jacobian
  //   products (VJPs):
  //
  //     grad_wrt_x = einsum("{eq_y},{eq_z}->{eq_x}", y, grad_wrt_z)
  //     grad_wrt_y = einsum("{eq_x},{eq_z}->{eq_y}", x, grad_wrt_z}
  //