In this function, the normal_eps is initially set based on the cell size. However, its value is normalized in the following line by the bound size. As a result, normal_eps becomes (bound_size * 2) times smaller than the cell size.
The resolutions of the hash encodings are always defined within the bounded regions, so is the normal_eps. The line you are referring is used to normalize the queries points, so they match the scale of normal_eps. I hope this helps.
In this function, the normal_eps is initially set based on the cell size. However, its value is normalized in the following line by the bound size. As a result, normal_eps becomes (bound_size * 2) times smaller than the cell size.
am I missing something?