mirror of
https://github.com/arsenetar/dupeguru.git
synced 2024-11-18 04:59:03 +00:00
109 lines
3.8 KiB
Python
109 lines
3.8 KiB
Python
# Created On: 2011/09/20
|
|
# Copyright 2011 Hardcoded Software (http://www.hardcoded.net)
|
|
#
|
|
# This software is licensed under the "BSD" License as described in the "LICENSE" file,
|
|
# which should be included with this package. The terms are also available at
|
|
# http://www.hardcoded.net/licenses/bsd_license
|
|
|
|
from hscommon.trans import tr as trbase
|
|
from hscommon.util import format_size
|
|
|
|
from core.app import (DupeGuru as DupeGuruBase, Column, format_time, format_timestamp,
|
|
format_perc, format_words, format_dupe_count, cmp_value)
|
|
from . import prioritize
|
|
|
|
tr = lambda s: trbase(s, 'columns')
|
|
|
|
class DupeGuru(DupeGuruBase):
|
|
COLUMNS = [
|
|
Column('name', tr("Filename")),
|
|
Column('folder_path', tr("Folder")),
|
|
Column('size', tr("Size (MB)")),
|
|
Column('duration', tr("Time")),
|
|
Column('bitrate', tr("Bitrate")),
|
|
Column('samplerate', tr("Sample Rate")),
|
|
Column('extension', tr("Kind")),
|
|
Column('mtime', tr("Modification")),
|
|
Column('title', tr("Title")),
|
|
Column('artist', tr("Artist")),
|
|
Column('album', tr("Album")),
|
|
Column('genre', tr("Genre")),
|
|
Column('year', tr("Year")),
|
|
Column('track', tr("Track Number")),
|
|
Column('comment', tr("Comment")),
|
|
Column('percentage', tr("Match %")),
|
|
Column('words', tr("Words Used")),
|
|
Column('dupe_count', tr("Dupe Count")),
|
|
]
|
|
DELTA_COLUMNS = {2, 3, 4, 5, 7}
|
|
METADATA_TO_READ = ['size', 'mtime', 'duration', 'bitrate', 'samplerate', 'title', 'artist',
|
|
'album', 'genre', 'year', 'track', 'comment']
|
|
MATCHPERC_COL = 15
|
|
DUPECOUNT_COL = 17
|
|
|
|
def __init__(self, view, appdata):
|
|
DupeGuruBase.__init__(self, view, appdata)
|
|
|
|
def _get_display_info(self, dupe, group, delta):
|
|
size = dupe.size
|
|
duration = dupe.duration
|
|
bitrate = dupe.bitrate
|
|
samplerate = dupe.samplerate
|
|
mtime = dupe.mtime
|
|
m = group.get_match_of(dupe)
|
|
if m:
|
|
percentage = m.percentage
|
|
dupe_count = 0
|
|
if delta:
|
|
r = group.ref
|
|
size -= r.size
|
|
duration -= r.duration
|
|
bitrate -= r.bitrate
|
|
samplerate -= r.samplerate
|
|
mtime -= r.mtime
|
|
else:
|
|
percentage = group.percentage
|
|
dupe_count = len(group.dupes)
|
|
return [
|
|
dupe.name,
|
|
str(dupe.folder_path),
|
|
format_size(size, 2, 2, False),
|
|
format_time(duration, with_hours=False),
|
|
str(bitrate),
|
|
str(samplerate),
|
|
dupe.extension,
|
|
format_timestamp(mtime,delta and m),
|
|
dupe.title,
|
|
dupe.artist,
|
|
dupe.album,
|
|
dupe.genre,
|
|
dupe.year,
|
|
str(dupe.track),
|
|
dupe.comment,
|
|
format_perc(percentage),
|
|
format_words(dupe.words) if hasattr(dupe, 'words') else '',
|
|
format_dupe_count(dupe_count)
|
|
]
|
|
|
|
def _get_dupe_sort_key(self, dupe, get_group, key, delta):
|
|
if key == self.MATCHPERC_COL:
|
|
m = get_group().get_match_of(dupe)
|
|
return m.percentage
|
|
if key == self.DUPECOUNT_COL:
|
|
return 0
|
|
r = cmp_value(getattr(dupe, self.COLUMNS[key].attr, ''))
|
|
if delta and (key in self.DELTA_COLUMNS):
|
|
r -= cmp_value(getattr(get_group().ref, self.COLUMNS[key].attr, ''))
|
|
return r
|
|
|
|
def _get_group_sort_key(self, group, key):
|
|
if key == self.MATCHPERC_COL:
|
|
return group.percentage
|
|
if key == self.DUPECOUNT_COL:
|
|
return len(group)
|
|
return cmp_value(getattr(group.ref, self.COLUMNS[key].attr, ''))
|
|
|
|
def _prioritization_categories(self):
|
|
return prioritize.all_categories()
|
|
|