export TEST_TMPDIR="${TFCI_MACOS_BAZEL_TEST_DIR_PATH}"
fi

# "TFCI_MACOS_INSTALL_BAZELISK_ENABLE" is used to decide if we need to install
# Bazelisk manually. We enable this for macOS x86 builds as those VMs do not
# have Bazelisk pre-installed. "TFCI_MACOS_INSTALL_BAZELISK_URL" contains the
# link to the Bazelisk binary which needs to be downloaded.
if [[ "${TFCI_MACOS_INSTALL_BAZELISK_ENABLE}" == 1 ]]; then

h5py >= 3.10.0
lit ~= 17.0.2
opt_einsum == 3.3.0
astunparse == 1.6.3
dill == 0.3.7
astor == 0.7.1
typing_extensions == 4.8.0
gast == 0.4.0
termcolor == 2.3.0
wrapt == 1.16.0
tblib == 2.0.0

# Install tensorboard, and keras
# Note that here we want the latest version that matches TF major.minor version
# Note that we must use nightly here as these are used in nightly jobs
# For release jobs, we will pin these on the release branch

if [[ "$TFCI_DOCKER_PULL_ENABLE" == 1 ]]; then
  # Simple retry logic for docker-pull errors. Sleeps for 15s if a pull fails.
  # Pulling an already-pulled container image will finish instantly, so
  # repeating the command costs nothing.
  docker pull "$TFCI_DOCKER_IMAGE" || sleep 15
  docker pull "$TFCI_DOCKER_IMAGE" || sleep 15
  docker pull "$TFCI_DOCKER_IMAGE"
fi 

    no longer include NCCL in the binary install. TensorFlow usage with multiple
    GPUs and NCCL requires upgrade to
    [NCCL 2.2](https://developer.nvidia.com/nccl). See updated install guides:
    [TensorFlow GPU support](https://www.tensorflow.org/install/gpu) and
    [Build TensorFlow from source](https://www.tensorflow.org/install/source).

  virtual AbstractTensorInterface* CreateTensor(
      DataType dtype, absl::Span<const int64_t> dim_sizes) = 0;

  typedef void (*MemoryReleaser)(void* data, size_t len, void* arg);

  // Create a tensor instance from the given data buffer and description.
  // `memory_releaser` will be called on destruction, and it's responsible for
  // cleaning up the underlying buffer.
  virtual AbstractTensorInterface* CreateTensor(

  TF_DeleteStatus(status);

  TFE_OpSetAttrType(op, "T", TFE_TensorHandleDataType(in));
  TFE_OpSetAttrInt(op, "group_size", group_size);
  TFE_OpSetAttrInt(op, "group_key", 123);
  TFE_OpSetAttrInt(op, "instance_key", 456);
  TFE_OpSetAttrString(op, "merge_op", "Add", 3);
  TFE_OpSetAttrString(op, "final_op", "Id", 2);
  std::vector<int64_t> subdiv_offsets;
  TFE_OpSetAttrIntList(op, "subdiv_offsets", subdiv_offsets.data(),

        "tf_private_symbols.lds",
    ],
)

genrule(
    name = "install_headers",
    srcs = [
        "//tensorflow/c:headers",
        "//tensorflow/c/eager:headers",
        "//tensorflow/cc:headers",
        "//tensorflow/core:headers",
        "@local_xla//xla/stream_executor:stream_executor_install_hdrs",
    ],
    outs = ["include"],
    cmd = """
    mkdir $@

  // Workaround for not being able to initialize empty map using {}
  std::unordered_set<string> empty_;
};

TEST_F(CApiFunctionTest, OneOp_ZeroInputs_OneOutput) {
  /*
   *                constant
   *                   |
   *                   v
   */
  // Define
  TF_Operation* c = ScalarConst(10, func_graph_, s_, "scalar10");
  Define(-1, {}, {}, {c}, {});

  // Use, run, and verify

@test "Wheel is installable" {
    python3 -m venv /tf/venv
    source /tf/venv/bin/activate
    python3 -m pip install --upgrade setuptools wheel
    python3 -m pip install "$TF_WHEEL"
}

@test "TensorFlow is importable" {
    source /tf/venv/bin/activate
    python3 -c 'import tensorflow as tf; t1=tf.constant([1,2,3,4]); t2=tf.constant([5,6,7,8]); print(tf.add(t1,t2).shape)'
}

# Is this still useful?
@test "TensorFlow has Keras" {

    choco install jdk8; \
    choco install KB2533623; \
    choco install KB2919355; \
    choco install KB2919442; \
    choco install KB2999226; \
    choco install KB3033929; \
    choco install KB3035131; \
    choco install maven; \
    choco install ninja; \
    choco install nodejs --version 9.3.0; \
    choco install nodejs.install --version 9.3.0; \