In the genotypes repertoire init function, a all-zero repertoire is created of the default dtype float. All following interactions with that repository silently convert genotypes to float.
If the genotypes are of another dtype like ints, this might induce problems down the line. For example, integers greater than 2**24 will not be stored correctly and very hard-to-debug problems ensue.
I don't see the reason why enforcing float dtype is necessary here.
In the genotypes repertoire init function, a all-zero repertoire is created of the default dtype float. All following interactions with that repository silently convert genotypes to float.
If the genotypes are of another dtype like ints, this might induce problems down the line. For example, integers greater than 2**24 will not be stored correctly and very hard-to-debug problems ensue.
I don't see the reason why enforcing float dtype is necessary here.