disabled give_readable_tensor_names for now, there's a bug

there are still some issues surrounding this regarding the int
tensors used to represent shapes, needs more debug

[Dockerfile] Deleted changed run_docker_jpk.sh, it is not necessary anymore because the work environment was changed

[Integration] Changed the position of the if condition to prevent the function building a unnecessary graph and shortened the custom node execution function by assuming that input0 is always v and input1 are always thresholds

[Integration] Interface of execute_custom_node.py to onnx_exec.py finished and first draft of unit test added