Allow copyright checker to check ANY non-excluded file

import re

import sys

import time

from subprocess import check_output

from pathlib import Path

from subprocess import PIPE

from subprocess import run

from typing import Dict

from typing import Iterable

from typing import Iterator

from typing import List

def check_file(path: str) -> bool:

def check_file(path: str, year: str) -> bool:

	"""

	Check for an up-to-date copyright notice in the first few lines of the given file

	"""

	with open(path) as file:

		lines = file.read(512)

	year = time.strftime('%Y')

	return bool(re.search(f'\\b(?:copyright)\\b.*\\b{year}\\b', lines, re.I))

def filter_excluded(paths: Iterable[Path]) -> List[Path]:

	"""

	Return a list of files from the input that are not excluded by git

	"""

	cmd = ['git', 'ls-files', '--other', '--cached', '--exclude-standard', '-z', '--']

	cmd.extend(path.as_posix() for path in paths)

	proc = run(cmd, stdout=PIPE, check=True)

	return list(split_paths(proc.stdout))

def get_file_years(paths: List[Path]) -> Dict[Path, str]:

	"""

	Return a mapping of paths to the year they where last changed (if they are tracked)

	"""

	output = dict()

	cmd = [

		'git', 'log', '--topo-order',

		'--format=format:%ad', '--date=format:%Y',

		'--name-only', '-z', '--',

	]

	cmd.extend(p.as_posix() for p in paths)

	proc = run(cmd, stdout=PIPE, check=True)

	regex = re.compile(br'(?P<year>[0-9]{4,})(?:\n|\r|\r\n)(?P<files>.*?)\x00\x00')

	for match in regex.finditer(proc.stdout):

		year = match.group('year').decode()

		for path in split_paths(match.group('files')):

			if path not in output:

				output[path] = year

	return output

def get_changed(paths: List[Path]) -> Iterator[Path]:

	"""

	Return an iterator of changed paths

	"""

	cmd = ['git', 'diff', '--no-commit-id', '--name-only', '-z', 'HEAD', '--']

	cmd.extend(path.as_posix() for path in paths)

	proc = run(cmd, stdout=PIPE, check=True)

	return split_paths(proc.stdout)

def split_paths(paths: bytes) -> Iterator[Path]:

	"""

	Return an iterator of Paths from a null-separated byte string list

	"""

	return (Path(p.decode()) for p in paths.split(b'\x00') if p != b'')

def main():

	"""

	CLI entrypoint

	"""

	cmd = ['git', 'diff', '--no-commit-id', '--name-only', '-z', 'HEAD']

	changed = set(check_output(cmd, text=True).split('\0'))

	changed.remove('')

	missing = []

	paths = filter_excluded(Path(arg) for arg in sys.argv[1:])

	for path in sys.argv[1:]:

		if path not in changed:

			continue

		if not check_file(path):

	year = time.strftime('%Y')

	years = {path: year for path in paths if path.is_file()}

	years.update(get_file_years(paths))

	years.update((path, year) for path in get_changed(paths))

	missing = []

	for path in paths:

		if path.is_file() and not check_file(path, years[path]):

			missing.append(path)

	if missing: