dustin
/
scripts
1
0
Fork 0
Code Issues Pull Requests Releases Activity

ripcd: Use chardet to detect CD-TEXT encoding

master
Dustin 2016-07-10 17:25:52 -05:00
parent 5dfd71b2fd
commit 3694f177df
1 changed files with 6 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
ripcd.py
View File

@ -3,6 +3,7 @@ from lxml import etree
import asyncio import asyncio
import aiohttp import aiohttp
import argparse import argparse
import chardet
import fnmatch import fnmatch
import functools import functools
import glob import glob
@ -25,7 +26,8 @@ class TocProtocol(asyncio.SubprocessProtocol):
self.done = asyncio.Future() self.done = asyncio.Future()
def pipe_data_received(self, fd, data): def pipe_data_received(self, fd, data):
for line in data.decode().splitlines(): detected = chardet.detect(data)
for line in data.decode(detected['encoding']).splitlines():
if not line: if not line:
continue continue
if line.startswith('Album title:'): if line.startswith('Album title:'):
@ -115,13 +117,14 @@ class Track(object):
basename = os.path.splitext(self.filename)[0] basename = os.path.splitext(self.filename)[0]
infname = '{}.inf'.format(basename) infname = '{}.inf'.format(basename)
try: try:
inf = open(infname) inf = open(infname, 'rb')
except OSError as e: except OSError as e:
sys.stderr.write('Could not read track info: {}\n'.format(e)) sys.stderr.write('Could not read track info: {}\n'.format(e))
return return
with inf: with inf:
for line in inf: for line in inf:
line = line.split('#')[0] detected = chardet.detect(line)
line = line.decode(detected['encoding']).split('#')[0]
if not line: if not line:
continue continue
try: try: