Optimize and parallelize structure compiling function

[StannisZhou]

Structure compiling takes 5.65s in https://github.com/NishanthJKumar/PGMax/blob/sanity_check/test_notebooks/sanity_check_optimize_mpbp_unpadded.py but 7.09s in https://github.com/NishanthJKumar/PGMax/blob/sanity_check/examples/sanity_check_example.py using the existing interface.

We should be able to optimize further, and parallelize the structure compiling process to significantly reduce the time taken.