I don’t know why floats aren’t included, but any float can be easily represented by an int with same bits, or a bytestring, using the struct module to convert between them, so there are clear workarounds.