poljakowski/sweeper
1
0
Fork 0
Code Issues Pull requests Projects Releases Wiki Activity

change default hash alg to sha1

Browse source
This commit is contained in:
darko-poljak 2014-02-08 20:27:30 +01:00
parent ea547d83c5
commit b41a4afa8b
1 changed files with 6 additions and 5 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

11
sweeper/sweeper.py
View file

@ -15,7 +15,7 @@ Options:
-b <blocksize>, --block-size=<blocksize> size of block used when reading -b <blocksize>, --block-size=<blocksize> size of block used when reading
file's content [default: 4096] file's content [default: 4096]
-d <hashalgs>, --digest-algs=<hashalgs> secure hash algorithm comma separated -d <hashalgs>, --digest-algs=<hashalgs> secure hash algorithm comma separated
list [default: md5] list [default: sha1]
note that multiple hashes will slow note that multiple hashes will slow
down sweeper down sweeper
-a <action>, --action=<action> action on duplicate files (pprint, -a <action>, --action=<action> action on duplicate files (pprint,
@ -71,6 +71,7 @@ from collections import defaultdict
from functools import partial from functools import partial
DEF_HASHALGS = ['sha1']
# some differences in python versions # some differences in python versions
# we prefer iter methods # we prefer iter methods
if sys.version_info[0] == 3: if sys.version_info[0] == 3:
@ -169,7 +170,7 @@ def _fbequal(fpath1, fpath2):
return False return False
def file_dups(topdirs=['./'], hashalgs=['md5'], block_size=4096, verbose=False, def file_dups(topdirs=['./'], hashalgs=DEF_HASHALGS, block_size=4096, verbose=False,
safe_mode=False): safe_mode=False):
"""Find duplicate files in directory list. Return directory """Find duplicate files in directory list. Return directory
with keys equal to file hash value and value as list of with keys equal to file hash value and value as list of
@ -237,7 +238,7 @@ def file_dups(topdirs=['./'], hashalgs=['md5'], block_size=4096, verbose=False,
return result return result
def iter_file_dups(topdirs=['./'], hashalgs=['md5'], block_size=4096, def iter_file_dups(topdirs=['./'], hashalgs=DEF_HASHALGS, block_size=4096,
safe_mode=False): safe_mode=False):
"""Find duplicate files in directory list. """Find duplicate files in directory list.
Yield tuple of file path, hash tuple and list of duplicate files Yield tuple of file path, hash tuple and list of duplicate files
@ -294,7 +295,7 @@ def _extract_files_for_action(topdirs, hashalgs, block_size, keep_prefix,
yield (files, result) yield (files, result)
def rm_file_dups(topdirs=['./'], hashalgs=['md5'], block_size=4096, def rm_file_dups(topdirs=['./'], hashalgs=DEF_HASHALGS, block_size=4096,
simulate=False, keep_prefix=None, verbose=False, simulate=False, keep_prefix=None, verbose=False,
safe_mode=False): safe_mode=False):
"""Remove duplicate files found in specified directory list. """Remove duplicate files found in specified directory list.
@ -318,7 +319,7 @@ def rm_file_dups(topdirs=['./'], hashalgs=['md5'], block_size=4096,
os.remove(f) os.remove(f)
def mv_file_dups(topdirs=['./'], hashalgs=['md5'], block_size=4096, def mv_file_dups(topdirs=['./'], hashalgs=DEF_HASHALGS, block_size=4096,
dest_dir='dups', simulate=False, keep_prefix=None, dest_dir='dups', simulate=False, keep_prefix=None,
verbose=False, safe_mode=False): verbose=False, safe_mode=False):
"""Move duplicate files found in specified directory list. """Move duplicate files found in specified directory list.