How to call tf.compat.v1 from the C++ API?

[Rabahjamal]

I have a frozen graph that was frozen in TensorFlow v1 from the python API, and I want to load it using TensorFlow v2.4 from the C++ API using something like tf.compat.v1 in python. I can do it in python with this script:

def wrap_frozen_graph(graph_def, inputs, outputs):
  def _imports_graph_def():
    tf.compat.v1.import_graph_def(graph_def, name="")
  wrapped_import = tf.compat.v1.wrap_function(_imports_graph_def, [])
  import_graph = wrapped_import.graph
  return wrapped_import.prune(
      tf.nest.map_structure(import_graph.as_graph_element, inputs),
      tf.nest.map_structure(import_graph.as_graph_element, outputs))


graph_def = tf.compat.v1.GraphDef()
loaded = graph_def.ParseFromString(open(path,'rb').read())


frozen_func = wrap_frozen_graph(graph_def=graph_def,
                                    inputs=["x:0"],
                                    outputs=["Identity:0", "Identity_1:0"])

But I didn't find any related document to do the same from the C++ API.

[anishlukk123]

is this a still a priortity?