*       v3-v4
         *
         */
        v1 = graph.addVertex(new ArtifactMetadata("g", "a1", "1.0"));
        graph.setEntry(v1);
        v2 = graph.addVertex(new ArtifactMetadata("g", "a2", "1.0"));
        v3 = graph.addVertex(new ArtifactMetadata("g", "a3", "1.0"));
        v4 = graph.addVertex(new ArtifactMetadata("g", "a4", "1.0"));

        // v1-->v2
        graph.addEdge(v1, v2, new MetadataGraphEdge("1.1", true, null, null, 2, 1));

 *
 * <pre>{@code
 * // Building a mutable value graph
 * MutableValueGraph<String, Double> graph =
 *     ValueGraphBuilder.undirected().allowsSelfLoops(true).build();
 * graph.putEdgeValue("San Francisco", "San Francisco", 0.0);
 * graph.putEdgeValue("San Jose", "San Jose", 0.0);
 * graph.putEdgeValue("San Francisco", "San Jose", 48.4);
 *
 * // Building an immutable value graph
 * ImmutableValueGraph<String, Double> immutableGraph =

 *
 * <h3>Usage</h3>
 *
 * Given an algorithm, for example:
 *
 * <pre>{@code
 * public <N> someGraphAlgorithm(N startNode, PredecessorsFunction<N> predecessorsFunction);
 * }</pre>
 *
 * you will invoke it depending on the graph representation you're using.
 *
 * <p>If you have an instance of one of the primary {@code common.graph} types ({@link Graph},
 * {@link ValueGraph}, and {@link Network}):

string DescribeCycle(const GraphCycles* cycles, const Graph& graph, int src,
                     int dst) {
  int32_t max_path_size = graph.num_node_ids() + 1;
  std::vector<int32> path(max_path_size);
  int32_t path_size = cycles->FindPath(dst, src, max_path_size, path.data());
  if (path_size == 0) {
    return "";
  }

  auto node_name = [&graph](int node_id) {
    if (!FastBoundsCheck(node_id, graph.num_node_ids())) {

import org.gradle.internal.serialize.graph.codecs.Decoding
import org.gradle.internal.serialize.graph.codecs.Encoding
import org.gradle.internal.serialize.graph.codecs.EncodingProducer
import org.gradle.internal.serialize.graph.ReadContext
import org.gradle.internal.serialize.graph.WriteContext
import org.gradle.internal.serialize.graph.readClassArray
import org.gradle.internal.serialize.graph.writeClassArray

  static Status Encapsulate(std::unique_ptr<Graph>* graph,
                            FunctionLibraryDefinition* flib_def);

  // b) we rewrite the function calls generated in phase (a) into XlaLaunch
  //    operators. We also convert the XlaClusterOutput output nodes of the
  //    function call into the outputs of the XlaLaunch operator.
  static Status BuildXlaLaunchOps(Graph* graph);

  struct XlaFunctionInfo {

 *
 * <h3>Usage</h3>
 *
 * Given an algorithm, for example:
 *
 * <pre>{@code
 * public <N> someGraphAlgorithm(N startNode, SuccessorsFunction<N> successorsFunction);
 * }</pre>
 *
 * you will invoke it depending on the graph representation you're using.
 *
 * <p>If you have an instance of one of the primary {@code common.graph} types ({@link Graph},
 * {@link ValueGraph}, and {@link Network}):

      VLOG(2) << "Run MLIR graph optimization pass: " << StringRefToView(name);
      VLOG(2) << "Graph #nodes " << (*graph)->num_nodes() << " #edges "
              << (*graph)->num_edges();
      timings.Reset({kTfMlirCategory, name.str()});
      pass_status = pass_registration.pass->Run(
          function_name, config_proto, *module_ref, **graph, *flib_def);
      timings.ReportAndStop();

  %graph = tf_executor.graph {
    tf_executor.fetch %arg0 : tensor<2xi32>
  }
  func.return %graph#0 : tensor<2xi32>
}

// Test with a TPUPartitionedCallOp that is not inlined.
func.func @dont_inline_tpu_partitioned_call(%arg0: tensor<2xi32> {tf._user_specified_name = "inputs_0"}, %arg1: tensor<2xi32> {tf._user_specified_name = "inputs_1"}) -> tensor<2xi32> {
  %graph = tf_executor.graph {

}

/*static*/ Status MarkForCompilationPassTestHelper::MarkForCompilation(
    std::unique_ptr<Graph>* graph,
    MarkForCompilationPassTestHelper::Options options) {
  FunctionDefLibrary flib;
  FunctionLibraryDefinition flib_def((*graph)->op_registry(), flib);
  return MarkForCompilation(graph, &flib_def, options);
}