Anyway to separate at whitespaces but avoid separating at file paths?

Question

I am trying to separate this string into a list using regex:

-y -hwaccel cuda -threads 8 -loglevel error -hide_banner -stats -i - -c:v hevc_nvenc -rc constqp -preset p7 -qp 18 C:\Users\User\Documents\Python\Smoothie\test 124\Resampled_vid.mp4

I am using the following method to separate it:

split(r'(?!\\)'+'\s+',f"{Settings[1]}".format(Input=InFile,Output=OutFile))

Output:
['-y', '-hwaccel', 'cuda', '-threads', '8', '-loglevel', 'error', '-hide_banner', '-stats', '-i', '-', '-c:v', 'hevc_nvenc', '-rc', 'constqp', '-preset', 'p7', '-qp', '18', 'C:\\Users\\User\\Documents\\Python\\Smoothie\\test', '124\\Resampled_vid.mp4']

Desired Output:

['-y', '-hwaccel', 'cuda', '-threads', '8', '-loglevel', 'error', '-hide_banner', '-stats', '-i', '-', '-c:v', 'hevc_nvenc', '-rc', 'constqp', '-preset', 'p7', '-qp', '18', 'C:\\Users\\User\\Documents\\Python\\Smoothie\\test 124\\Resampled_vid.mp4']

Is there anyway, I can exclusively avoid splitting at a file path?

Answer 1

I would use an re.findall approach here:

inp = "-y -hwaccel cuda -threads 8 -loglevel error -hide_banner -stats -i - -c:v hevc_nvenc -rc constqp -preset p7 -qp 18 C:\Users\User\Documents\Python\Smoothie\test 124\Resampled_vid.mp4"
parts = re.findall(r'[A-Z]+:(?:\\[^\\]+)+\.\w+|\S+', inp)
print(parts)

['-y', '-hwaccel', 'cuda', '-threads', '8', '-loglevel', 'error', '-hide_banner',
 '-stats', '-i', '-', '-c:v', 'hevc_nvenc', '-rc', 'constqp', '-preset', 'p7',
 '-qp', '18',
 'C:\\Users\\User\\Documents\\Python\\Smoothie\test 124\\Resampled_vid.mp4']

The regex pattern used here says to match, alternatively:

[A-Z]+:(?:\\[^\\]+)+\.\w+  a file path
|                          OR
\S+                        any group of non whitespace characters

The trick here is to eagerly try to match a file path first. Only that failing do we try to match one word/term at a time.