This feels like it should be very simple, but I haven't been able to find an answer..
In a python script I am reading in data from a USB device (x and y movements of a USB mouse). it arrives in single ASCII characters. I can easily convert to unsigned integers (0-255) using ord. But, I would like it as signed integers (-128 to 127) - how can I do this?
Any help greatly appreciated! Thanks a lot.
Subtract 256 if over 127:
unsigned = ord(character)
signed = unsigned - 256 if unsigned > 127 else unsigned
Alternatively, repack the byte with the struct module:
from struct import pack, unpack
signed = unpack('B', pack('b', unsigned))[0]
or directly from the character:
signed = unpack('B', character)[0]
from ctypes import c_int8
value = c_int8(191).value
use ctypes with your ord() value - should be -65 in this case
ex. from string data
from ctypes import c_int8
data ='BF'
value1 = int(data, 16) # or ord(data.decode('hex'))
value2 = c_int8(value1).value
value1 is 16bit integer representation of hex 'BF' and value2 is 8bit representation
I know this is an old question, but I haven't found a satisfying answer elsewhere.
You can use the array module (with the extra convenience that it converts complete buffers):
from array import array
buf = b'\x00\x01\xff\xfe'
print(array('b', buf))
# result: array('b', [0, 1, -1, -2])
Use this function to get a signed 8bit integer value
def to8bitSigned(num):
mask7 = 128 #Check 8th bit ~ 2^8
mask2s = 127 # Keep first 7 bits
if (mask7 & num == 128): #Check Sign (8th bit)
num = -((~int(num) + 1) & mask2s) #2's complement
return num
To convert any input bytes into signed integers:
def signed8bit_to_int(input):
(((input >> 7) * 128) ^ input) - ((input >> 7) * 128)
Examples:
signed8bit_to_int(0xc0) = -64
signed8bit_to_int(0xbf) = -65
signed8bit_to_int(0x0f) = 15
Explanation with 0xC0 as an example:
The question doesn't clearly state if the single character is given as str or bytes.
The following answer is especially useful, if the input is an instance of bytes:
Since Python 3.2 there is a class method int.from_bytes:
int.from_bytes(b'\x81', byteorder='big', signed=True) # -127
int.from_bytes(b'a', byteorder='big', signed=True) # 97
For 16-bit integers, one would simply pass 2 bytes instead of 1:
int.from_bytes(b'\x81\x80', byteorder='big', signed=True) # -32384
int.from_bytes(b'ab', byteorder='big', signed=True) # 24930
Of course, you may need to specify byteorder='little'.
However, if the input is an ASCII character (0 - 128), I don't understand why you want a signed integer, because the result for your input will always be positive anyway (if I'm not mistaken). Python's integers use a flexible amount of bytes.