ModuleNotFoundError: No module named 'numpy.core.multiarray\r' - python-3.x

Question

I am trying to load a pkl file,
pkl_file = open(sys.argv[1], 'rb')
world = pickle.load(pkl_file)
but I get an error from these lines
Traceback (most recent call last):
File "E:/python/test.py", line 186, in <module>
world = pickle.load(pkl_file)
ModuleNotFoundError: No module named 'numpy.core.multiarray\r'
I am using Windows 10, python 3.7, and installed four packages (numpy 1.17.2, opencv-python 4.1.1.26, pip 19.2.3, setuptools 41.2.0 ). I have tried to change "rb" to "r", but still got the error, how can I fix this?

Answer 1

I think there are two problems here.
First, your pickle is or contains a NumPy object, which is not part of the standard library. Therefore you must ensure that NumPy is installed into your current Python environment and imported before you try to load the pickled object. Depending on your setup, installation may be as simple as,
pip install numpy
Then you must add the line,
import numpy as np
to the top of your script.
Second, it looks like Python is encountering this issue, where your binary file was erroneously saved as text on Windows, resulting in resulted in each '\n' being converted to '\r\n'. To fix this, you must re-convert to '\r\n' back to '\n'. So long as the file isn't huge, this usually isn't very painful.
Here is a relatively complete example:
import sys
import numpy as np
src = sys.argv[1] # path to your file
data = open(src).read().replace('\r\n', '\n') # read and replace file contents
dst = src + ".tmp"
open(dst, "w").write(data) # save a temporary file
world = pickle.load(open(dst, "rb"), encoding='latin1')

Answer 2

Ok, I just had to figure this out for myself, and I solved it. All you have to do is change all the "\r\n" to "\n". You can do this in multiple ways. You can go into Notepad++ and change line endings from CR LF to just LF. Or programmatically you can do
open(newfile, 'w', newline = '\n').write(open(oldfile, 'r').read())