adelgado/find_duplicate_files
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

add output file

Browse source
This commit is contained in:
Antonio J. Delgado 2023-01-26 09:14:01 +02:00
parent 251c3078e0
commit 51517f5212
1 changed files with 13 additions and 9 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

10
find_duplicate_files/find_duplicate_files.py
View file

@ -17,7 +17,7 @@ import re
class find_duplicate_files: class find_duplicate_files:
def __init__(self, debug_level, log_file, dummy, first_directory, second_directory, exclude, limit): def __init__(self, debug_level, log_file, dummy, first_directory, second_directory, exclude, limit, output_file):
''' Initial function called when object is created ''' ''' Initial function called when object is created '''
self.config = dict() self.config = dict()
self.config['debug_level'] = debug_level self.config['debug_level'] = debug_level
@ -31,6 +31,7 @@ class find_duplicate_files:
self.second_directory = second_directory self.second_directory = second_directory
self.exclude = exclude self.exclude = exclude
self.limit = limit self.limit = limit
self.output_file = output_file
self._init_db_cache() self._init_db_cache()
@ -41,12 +42,14 @@ class find_duplicate_files:
total = len(first_files) total = len(first_files)
count = 0 count = 0
with open(self.output_file, 'w') as output_pointer:
for hash in first_files: for hash in first_files:
count += 1 count += 1
self._log.info(f"# Checking file {count} of {total}") self._log.info(f"# Checking file {count} of {total}")
if hash in second_files: if hash in second_files:
self._log.info(f"#File '{first_files[hash]}' is dupe with '{second_files[hash]}'.") self._log.info(f"#File '{first_files[hash]}' is dupe with '{second_files[hash]}'.")
self._log.info(f"rm '{first_files[hash]}'") self._log.info(f"rm '{first_files[hash]}'")
output_pointer.write(f"rm '{first_files[hash]}'\n")
def _init_db_cache(self, cache_file='/var/cache/find_duplicate_files.cache.sql'): def _init_db_cache(self, cache_file='/var/cache/find_duplicate_files.cache.sql'):
self.cache_file = cache_file self.cache_file = cache_file
@ -161,9 +164,10 @@ class find_duplicate_files:
@click.option('--second-directory', '-s', required=True, help='Second directory to find files') @click.option('--second-directory', '-s', required=True, help='Second directory to find files')
@click.option('--exclude', '-e', multiple=True, help='Regular expression pattern to exclude from files and directories.') @click.option('--exclude', '-e', multiple=True, help='Regular expression pattern to exclude from files and directories.')
@click.option('--limit', '-l', default=0, type=int, help='Limit to a certain number of files to check.') @click.option('--limit', '-l', default=0, type=int, help='Limit to a certain number of files to check.')
@click.option('--output-file', '-o', default='/tmp/delete_duplicates_commands.sh', help='File to write the commands to delete duplicate files. USE WITH CAUTION!')
@click_config_file.configuration_option() @click_config_file.configuration_option()
def __main__(debug_level, log_file, dummy, first_directory, second_directory, exclude, limit): def __main__(debug_level, log_file, dummy, first_directory, second_directory, exclude, limit, output_file):
return find_duplicate_files(debug_level, log_file, dummy, first_directory, second_directory, exclude, limit) return find_duplicate_files(debug_level, log_file, dummy, first_directory, second_directory, exclude, limit, output_file)
if __name__ == "__main__": if __name__ == "__main__":
__main__() __main__()