X-Git-Url: https://wannabe.guru.org/gitweb/?a=blobdiff_plain;f=directory_filter.py;h=b076badf25dff7e34e2358b3e45d52c49692a424;hb=31c81f6539969a5eba864d3305f9fb7bf716a367;hp=03602d156a0929ba6b60ac094d264359e3426d6d;hpb=52290f9c9e0eeaba3d5a067043f5ba98c9b386e5;p=python_utils.git diff --git a/directory_filter.py b/directory_filter.py index 03602d1..b076bad 100644 --- a/directory_filter.py +++ b/directory_filter.py @@ -3,7 +3,7 @@ import hashlib import logging import os -from typing import Any, Optional +from typing import Any, Dict, Optional, Set logger = logging.getLogger(__name__) @@ -30,14 +30,16 @@ class DirectoryFileFilter(object): >>> os.remove(testfile) """ + def __init__(self, directory: str): super().__init__() import file_utils + if not file_utils.does_directory_exist(directory): raise ValueError(directory) self.directory = directory - self.md5_by_filename = {} - self.mtime_by_filename = {} + self.md5_by_filename: Dict[str, str] = {} + self.mtime_by_filename: Dict[str, float] = {} self._update() def _update(self): @@ -49,9 +51,11 @@ class DirectoryFileFilter(object): def _update_file(self, filename: str, mtime: Optional[float] = None): import file_utils + assert file_utils.does_file_exist(filename) if mtime is None: mtime = file_utils.get_file_raw_mtime(filename) + assert mtime if self.mtime_by_filename.get(filename, 0) != mtime: md5 = file_utils.get_file_md5(filename) logger.debug(f'Computed/stored {filename}\'s MD5 at ts={mtime} ({md5})') @@ -97,22 +101,26 @@ class DirectoryAllFilesFilter(DirectoryFileFilter): >>> os.remove(testfile) """ + def __init__(self, directory: str): - self.all_md5s = set() + self.all_md5s: Set[str] = set() super().__init__(directory) def _update_file(self, filename: str, mtime: Optional[float] = None): import file_utils + assert file_utils.does_file_exist(filename) if mtime is None: mtime = file_utils.get_file_raw_mtime(filename) + assert mtime if self.mtime_by_filename.get(filename, 0) != mtime: md5 = file_utils.get_file_md5(filename) self.mtime_by_filename[filename] = mtime self.md5_by_filename[filename] = md5 self.all_md5s.add(md5) - def apply(self, item: Any) -> bool: + def apply(self, item: Any, ignored_filename: str = None) -> bool: + assert not ignored_filename self._update() mem_hash = hashlib.md5() mem_hash.update(item) @@ -122,4 +130,5 @@ class DirectoryAllFilesFilter(DirectoryFileFilter): if __name__ == '__main__': import doctest + doctest.testmod()