lkj Recipes

[thorwhalen]

A bunch of recipes using lkj tools.

[thorwhalen]

Get a list of unique relative file paths

A unique_relative_paths function can be particularly useful when working with collections of files that may have duplicate filenames located in different directories. In scenarios such as generating concise and human-readable labels for files, creating summaries, or displaying file paths in user interfaces or logs, it’s important to present paths that are both unique and as short as possible. By intelligently resolving name collisions through the inclusion of additional parent directories only when necessary, unique_relative_paths ensures that each file can be distinctly identified without overwhelming the user with full absolute paths. This function is especially beneficial in large projects or datasets where duplicate filenames are common, helping to maintain clarity and prevent confusion caused by ambiguous file references.

from functools import partial
import os

unique_relative_paths = partial(unique_affixes, suffix=True, ingress=partial(str.split, sep=os.path.sep), egress=os.path.sep.join)

# Example

test_paths = paths = [
    '/home/user/project/unique/file.py',
    '/home/user/project/src/file1.py',
    '/home/user/project/tests/file1.py',
    '/home/user/project/src/utils/file2.py',
    '/home/user/project/docs/file2.py'
]
assert unique_relative_paths(test_paths) == [
    'file.py',
    'src/file1.py',
    'tests/file1.py',
    'utils/file2.py',
    'docs/file2.py',
]

Note that ingress=partial(str.split, sep=os.path.sep) is os aware as far as path separator goes, but for a even more robust way (e.g. handling drive letters, etc.) you can do ingress=lambda p: pathlib.Path(p).parts, or if you need it to be picklable,
dol.Pipe(pathlib.Path, operator.attrgetter('parts')).