1
0
mirror of https://github.com/janeczku/calibre-web synced 2024-10-31 23:26:20 +00:00
calibre-web/cps/comic.py

169 lines
6.2 KiB
Python
Raw Normal View History

# -*- coding: utf-8 -*-
# This file is part of the Calibre-Web (https://github.com/janeczku/calibre-web)
# Copyright (C) 2018 OzzieIsaacs
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
from __future__ import division, print_function, unicode_literals
import os
import io
from . import logger, isoLanguages
from .constants import BookMeta
try:
from PIL import Image as PILImage
use_PIL = True
except ImportError as e:
use_PIL = False
log = logger.create()
try:
from comicapi.comicarchive import ComicArchive, MetaDataStyle
use_comic_meta = True
try:
from comicapi import __version__ as comic_version
except (ImportError):
comic_version = ''
except ImportError as e:
log.debug('Cannot import comicapi, extracting comic metadata will not work: %s', e)
import zipfile
import tarfile
try:
import rarfile
use_rarfile = True
except ImportError as e:
log.debug('Cannot import rarfile, extracting cover files from rar files will not work: %s', e)
use_rarfile = False
use_comic_meta = False
def _cover_processing(tmp_file_name, img, extension):
if use_PIL:
# convert to jpg because calibre only supports jpg
if extension in ('.png', '.webp'):
imgc = PILImage.open(io.BytesIO(img))
im = imgc.convert('RGB')
tmp_bytesio = io.BytesIO()
im.save(tmp_bytesio, format='JPEG')
img = tmp_bytesio.getvalue()
prefix = os.path.dirname(tmp_file_name)
if img:
tmp_cover_name = prefix + '/cover.jpg'
image = open(tmp_cover_name, 'wb')
image.write(img)
image.close()
else:
tmp_cover_name = None
return tmp_cover_name
def _extractCover(tmp_file_name, original_file_extension, rarExceutable):
cover_data = extension = None
if use_comic_meta:
archive = ComicArchive(tmp_file_name)
2019-08-03 12:09:54 +00:00
for index, name in enumerate(archive.getPageNameList()):
ext = os.path.splitext(name)
if len(ext) > 1:
extension = ext[1].lower()
if extension in ('.jpg', '.jpeg', '.png', '.webp'):
2019-08-03 12:09:54 +00:00
cover_data = archive.getPage(index)
break
else:
if original_file_extension.upper() == '.CBZ':
cf = zipfile.ZipFile(tmp_file_name)
for name in cf.namelist():
ext = os.path.splitext(name)
if len(ext) > 1:
extension = ext[1].lower()
if extension in ('.jpg', '.jpeg', '.png', '.webp'):
cover_data = cf.read(name)
break
elif original_file_extension.upper() == '.CBT':
cf = tarfile.TarFile(tmp_file_name)
for name in cf.getnames():
ext = os.path.splitext(name)
if len(ext) > 1:
extension = ext[1].lower()
if extension in ('.jpg', '.jpeg', '.png', '.webp'):
cover_data = cf.extractfile(name).read()
break
elif original_file_extension.upper() == '.CBR' and use_rarfile:
try:
rarfile.UNRAR_TOOL = rarExceutable
cf = rarfile.RarFile(tmp_file_name)
for name in cf.getnames():
ext = os.path.splitext(name)
if len(ext) > 1:
extension = ext[1].lower()
if extension in ('.jpg', '.jpeg', '.png', '.webp'):
cover_data = cf.read(name)
break
except Exception as e:
log.debug('Rarfile failed with error: %s', e)
return _cover_processing(tmp_file_name, cover_data, extension)
def get_comic_info(tmp_file_path, original_file_name, original_file_extension, rarExceutable):
if use_comic_meta:
archive = ComicArchive(tmp_file_path, rar_exe_path=rarExceutable)
if archive.seemsToBeAComicArchive():
if archive.hasMetadata(MetaDataStyle.CIX):
style = MetaDataStyle.CIX
elif archive.hasMetadata(MetaDataStyle.CBI):
style = MetaDataStyle.CBI
else:
style = None
2019-08-03 12:09:54 +00:00
# if style is not None:
loadedMetadata = archive.readMetadata(style)
2019-08-03 12:09:54 +00:00
lang = loadedMetadata.language
if lang:
if len(lang) == 2:
loadedMetadata.language = isoLanguages.get(part1=lang).name
elif len(lang) == 3:
loadedMetadata.language = isoLanguages.get(part3=lang).name
else:
loadedMetadata.language = ""
return BookMeta(
file_path=tmp_file_path,
extension=original_file_extension,
title=loadedMetadata.title or original_file_name,
author=" & ".join([credit["person"] for credit in loadedMetadata.credits if credit["role"] == "Writer"]) or u'Unknown',
cover=_extractCover(tmp_file_path, original_file_extension, rarExceutable),
description=loadedMetadata.comments or "",
tags="",
series=loadedMetadata.series or "",
series_id=loadedMetadata.issue or "",
languages=loadedMetadata.language)
return BookMeta(
file_path=tmp_file_path,
extension=original_file_extension,
title=original_file_name,
author=u'Unknown',
cover=_extractCover(tmp_file_path, original_file_extension, rarExceutable),
description="",
tags="",
series="",
series_id="",
languages="")