vertex_percolation returns unexpected sizes
According to the document, the size
returned from vertex_percolation should be
Size of the largest component prior to removal of each vertex.
That is, the size should at least begin with 1
while there's only one vertex left. [1, ...]
.
However, it actually always returns larger than 1
.
After reviewing the backend code, I found that the size actually includes neighbors. This doesn't match the inplementation in the reference paper
M. E. J. Newman, R. M. Ziff, “A fast Monte Carlo algorithm for site or bond percolation”, Phys. Rev. E 64, 016706 (2001) DOI: 10.1103/PhysRevE.64.016706, arXiv: cond-mat/0101295