I'm currently working on a project which requires me to modify bit values of float variables. To do this, I am using the struct package to acquire the binary values of a float as a string, like so:
num = -5.661166216897089
s = struct.pack('!f', num)
b = ''.join(format(c, '08b')for c in s)
I will then modify the desired bits with splicing. My question is:
After modifying the string of bits as needed, how can I use it to create a new float value with the modified bit sequence? Defining it as such:
new_float = 0b11000000101101010010100001000110
does not work, I suspect as it is being interpreted as an integer. Is there a way to force float creation?
Thanks ^^
From b you create a new str, v by changing some bits of b. For example:
v='11001110101101010010100001000110'
Now, you can convert that to bytes:
v_bytes = bytes([int(v[i:i+8], 2) for i in range(4)])
Which looks like:
print(v_bytes)
b'\xce\x9d:u'
and that can be given to struct.unpack:
new_num = struct.unpack("!f", v_bytes)[0]
unpack returns a tuple, that's why you must take its first element with [0].