graph_->AddEdge(input_node_it->second, 0, dst_node, dst_index);
  return absl::OkStatus();
}

Status Exporter::AddEdge(Operation* inst) {
  // For tf_executor.fetch, add only its data edges. Control edges are captured
  // later.
  if (auto fetch = llvm::dyn_cast<mlir::tf_executor::FetchOp>(inst)) {
    for (auto operand_and_idx : llvm::enumerate(fetch.getOperands())) {
      Value operand = operand_and_idx.value();

  graph_->AddEdge(input_node_it->second, 0, dst_node, dst_index);
  return absl::OkStatus();
}

Status Exporter::AddEdge(Operation* inst) {
  // For tf_executor.fetch, add only its data edges. Control edges are captured
  // later.
  if (auto fetch = llvm::dyn_cast<mlir::tf_executor::FetchOp>(inst)) {
    for (auto operand_and_idx : llvm::enumerate(fetch.getOperands())) {
      Value operand = operand_and_idx.value();

			*em = (*em)[:len(*em)-1]
			return
		}
	}
}

// Edge contains any attributes to be represented about edges in a graph.
type Edge struct {
	Src, Dest *Node
	// The summary weight of the edge
	Weight, WeightDiv int64

	// residual edges connect nodes that were connected through a
	// separate node, which has been removed from the report.
	Residual bool

  // Check for undeclared outputs before Encapsulation, so we can give a better
  // error message.
  // TODO(phawkins): merge this with the encapsulation code to avoid the extra
  // O(n) pass over the edges.
  for (const Edge* e : (*graph)->edges()) {
    if (!e->IsControlEdge() &&
        e->src()->attrs().Find(kXlaClusterIdAttr) != nullptr &&
        e->dst()->attrs().Find(kXlaClusterIdAttr) == nullptr &&

        }

        private visitEdges(NodeBuilder node, Set<NodeBuilder> seenNodes, Set<EdgeBuilder> edges) {
            for (EdgeBuilder edge : node.deps) {
                edges.add(edge)
                if (seenNodes.add(edge.selected)) {
                    visitEdges(edge.selected, seenNodes, edges)
                }
            }
        }

        /**

                << (*graph)->num_edges();
        is_module_updated = true;
      }
    } else if (pass_state == MlirOptimizationPassState::FallbackEnabled) {
      VLOG(2) << "Run MLIR graph optimization pass with fallback: "
              << StringRefToView(name);
      VLOG(2) << "Graph #nodes " << (*graph)->num_nodes() << " #edges "
              << (*graph)->num_edges();

			M[obj] = &graphNode{obj: obj}
		}
	}

	// compute edges for graph M
	// (We need to include all nodes, even isolated ones, because they still need
	// to be scheduled for initialization in correct order relative to other nodes.)
	for obj, n := range M {
		// for each dependency obj -> d (= deps[i]), create graph edges n->s and s->n
		for d := range objMap[obj].deps {

			M[obj] = &graphNode{obj: obj}
		}
	}

	// compute edges for graph M
	// (We need to include all nodes, even isolated ones, because they still need
	// to be scheduled for initialization in correct order relative to other nodes.)
	for obj, n := range M {
		// for each dependency obj -> d (= deps[i]), create graph edges n->s and s->n
		for d := range objMap[obj].deps {

    auto const& pair = visited.insert(nout.node());
    if (pair.second) {
      queue.push_back(std::make_pair(nout.node(), static_cast<Node*>(nullptr)));
    }
  }
  // BFS from nodes in 'inputs_' along out edges for the entire graph. Internal
  // output nodes are recorded during the traversal. All nodes that are output
  // nodes but not internal output nodes are considered the frontier of the

  ASSERT_EQ(f1_rewritten->signature().output_arg_size(), 1);
  EXPECT_EQ(f1_rewritten->signature().output_arg(0).type(), DT_BOOL);

  // Check node "if" input and output edges.
  auto node_name_index = g->BuildNodeNameIndex();
  const Node *if_node = node_name_index.at("if");
  ASSERT_NE(if_node, nullptr);
  const Node *input_node;
  TF_CHECK_OK(if_node->input_node(1, &input_node));