Cutlery/calibre
1
0
Fork 0
mirror of https://github.com/kovidgoyal/calibre.git synced 2025-07-07 10:14:46 -04:00
Code Issues Packages Projects Releases Wiki Activity

Fix #1712 (hopefully). Make cover detection significantly more robust.

Browse Source
This commit is contained in:
Marshall T. Vandegrift 2009-01-27 19:56:39 -05:00
parent 0e6a55fb7e
commit c2aa71e851
1 changed files with 44 additions and 26 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

68
src/calibre/ebooks/oeb/base.py
View File

@ -796,6 +796,10 @@ class TOC(object):
class OEBBook(object): class OEBBook(object):
COVER_SVG_XP = XPath('h:body//svg:svg[position() = 1]')
COVER_OBJECT_XP = XPath('h:body//h:object[@data][position() = 1]')
COVER_IMG_XP = XPath('h:body//h:img[@src][position() = 1]')
def __init__(self, opfpath=None, container=None, encoding=None, def __init__(self, opfpath=None, container=None, encoding=None,
logger=FauxLogger()): logger=FauxLogger()):
if opfpath and not container: if opfpath and not container:
@ -971,7 +975,7 @@ class OEBBook(object):
ncx = item.data ncx = item.data
self.manifest.remove(item) self.manifest.remove(item)
title = xpath(ncx, 'ncx:docTitle/ncx:text/text()') title = xpath(ncx, 'ncx:docTitle/ncx:text/text()')
title = title[0].strip() if title else unicode(self.metadata.title) title = title[0].strip() if title else unicode(self.metadata.title[0])
self.toc = toc = TOC(title) self.toc = toc = TOC(title)
navmaps = xpath(ncx, 'ncx:navMap') navmaps = xpath(ncx, 'ncx:navMap')
for navmap in navmaps: for navmap in navmaps:
@ -1051,41 +1055,55 @@ class OEBBook(object):
if self._toc_from_html(opf): return if self._toc_from_html(opf): return
self._toc_from_spine(opf) self._toc_from_spine(opf)
def _ensure_cover_image(self): def _locate_cover_image(self):
cover = None if self.metadata.cover:
id = str(self.metadata.cover[0])
item = self.manifest.ids.get(id, None)
if item is not None:
return item
hcover = self.spine[0] hcover = self.spine[0]
if 'cover' in self.guide: if 'cover' in self.guide:
href = self.guide['cover'].href href = self.guide['cover'].href
item = self.manifest.hrefs[href] item = self.manifest.hrefs[href]
media_type = item.media_type media_type = item.media_type
if media_type in OEB_RASTER_IMAGES: if media_type in OEB_IMAGES:
cover = item return item
elif media_type in OEB_DOCS: elif media_type in OEB_DOCS:
hcover = item hcover = item
html = hcover.data html = hcover.data
if cover is not None: if MS_COVER_TYPE in self.guide:
pass
elif self.metadata.cover:
id = str(self.metadata.cover[0])
cover = self.manifest.ids[id]
elif MS_COVER_TYPE in self.guide:
href = self.guide[MS_COVER_TYPE].href href = self.guide[MS_COVER_TYPE].href
cover = self.manifest.hrefs[href] item = self.manifest.hrefs.get(href, None)
elif xpath(html, '//h:img[position()=1]'): if item is not None and item.media_type in OEB_IMAGES:
img = xpath(html, '//h:img[position()=1]')[0] return item
href = hcover.abshref(img.get('src')) if self.COVER_SVG_XP(html):
cover = self.manifest.hrefs[href] svg = copy.deepcopy(self.COVER_SVG_XP(html)[0])
elif xpath(html, '//h:object[position()=1]'):
object = xpath(html, '//h:object[position()=1]')[0]
href = hcover.abshref(object.get('data'))
cover = self.manifest.hrefs[href]
elif xpath(html, '//svg:svg[position()=1]'):
svg = copy.deepcopy(xpath(html, '//svg:svg[position()=1]')[0])
href = os.path.splitext(hcover.href)[0] + '.svg' href = os.path.splitext(hcover.href)[0] + '.svg'
id, href = self.manifest.generate(hcover.id, href) id, href = self.manifest.generate(hcover.id, href)
cover = self.manifest.add(id, href, SVG_MIME, data=svg) item = self.manifest.add(id, href, SVG_MIME, data=svg)
if cover and not self.metadata.cover: return item
self.metadata.add('cover', cover.id) if self.COVER_OBJECT_XP(html):
object = self.COVER_OBJECT_XP(html)[0]
href = hcover.abshref(object.get('data'))
item = self.manifest.hrefs.get(href, None)
if item is not None and item.media_type in OEB_IMAGES:
return item
if self.COVER_IMG_XP(html):
img = self.COVER_IMG_XP(html)[0]
href = hcover.abshref(img.get('src'))
item = self.manifest.hrefs.get(href, None)
if item is not None and item.media_type in OEB_IMAGES:
return item
return None
def _ensure_cover_image(self):
cover = self._locate_cover_image()
if not cover:
return
if self.metadata.cover:
self.metadata.cover[0].value = cover.id
return
self.metadata.add('cover', cover.id)
def _all_from_opf(self, opf): def _all_from_opf(self, opf):
self._metadata_from_opf(opf) self._metadata_from_opf(opf)