Read/Write binary file - Printable Version

Read/Write binary file - Printable Version

+- Python Forum (https://python-forum.io)
+-- Forum: Python Coding (https://python-forum.io/forum-7.html)
+--- Forum: General Coding Help (https://python-forum.io/forum-8.html)
+--- Thread: Read/Write binary file (/thread-32272.html)

Read/Write binary file - deanhystad - Jan-31-2021

I have a binary file that contains 32 bit floats. I would like to read the file into a matrix, rotate the matrix, and then write the matrix to a file.

In the original file the values would be arranged like this:
a0, b0, a1, b1, a2, b2, a3, b3...aN, bN
where a,b can be thought of as rows and 1, 2, 3 as columns.

In the resulting file the values would be arranged like this:
a0, a1, a2, a3...aN, b0, b1, b2, b3...bN

I'm beginning to think I might need to use the array library to read and write the files, and write some C code to do the rotation. Does anyone know of other ways?

RE: Read/Write binary file - Larz60+ - Feb-01-2021

I don't think it's nearly as complicated as you think.
Please supply a sample file.

RE: Read/Write binary file - deanhystad - Feb-01-2021

I don't know how to provide a sample file, so here is some code to generate the sample file and demonstrate the kind of manipulation I need.

import numpy as np
from array import array

a = array('f', [x for _ in range(10) for x in range(4)])
with open('sample_order.bin', 'wb') as file:
    a.tofile(file)

a = array('f')
with open('sample_order.bin', 'rb') as file:
    a.fromfile(file, 40)
print('Sample order file', a, sep='\n')

b = np.array(a)
print('As numpy array', b, sep='\n')
c = b.reshape((10, 4))
print('As a matrix', c, sep='\n')
d = c.transpose()
print('As matrix in channel order', d, sep='\n')
e = array('f', d.flatten())
print('Channel order file', e, sep='\n')

with open('channel_order.bin', 'wb') as file:
    e.tofile(file)

I start with an array like this:
0,1,2,3,0,1,2,3,0,1,2,3,0,1,2,3,0,1,2,3,0,1,2,3,0,1,2,3,0,1,2,3,0,1,2,3,0,1,2,3

I want to convert this to:
0,0,0,0,0,0,0,0,0,0,1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,2,2,3,3,3,3,3,3,3,3,3,3

I am doing this by turning it into a matrix, transposing the matrix and then flattening the matrix, I know there has to be a better way.

RE: Read/Write binary file - Larz60+ - Feb-01-2021

for 2d array, I'd use something like below

But you have a single dimension array, so you're way ahead of me on this.

import numpy as np
from scipy.ndimage.interpolation import rotate
import os

# make sure file can be found in script directory
os.chdir(os.path.abspath(os.path.dirname(__file__)))

def rotate_and_save(infilename, outfilename, rotation_angle_degrees=90):    
    orig_array = np.fromfile(infilename, dtype=np.int64)
    print(f"\noriginal array\n{orig_array}")
    new_array = rotate(orig_array, angle=rotation_angle_degrees)
    print(f"\nnew array\n{new_array}")
    np.tofile(outfilename, dtype=np.int64)

def main():
    # Use any desired angle
    rotate_and_save('channel_order.bin', 'rotated_channel_order.bin', 
        rotation_angle_degrees=180)

if __name__ == '__main__':
    main(