Cutlery/calibre
1
0
Fork 0
mirror of https://github.com/kovidgoyal/calibre.git synced 2025-07-09 03:04:10 -04:00
Code Issues Packages Projects Releases Wiki Activity

EPUB metadata: Extract the cover image from the html it is embededd in if possible, instead of rendering the html. Removes the white margins on covers and speeds up cover extraction

Browse Source
This commit is contained in:
Kovid Goyal 2010-02-07 12:49:54 -07:00
parent 6a24c74e50
commit a1365513dd
1 changed files with 18 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
src/calibre/ebooks/__init__.py
View File

@ -70,6 +70,19 @@ def extract_cover_from_embedded_svg(html, base, log):
if href and os.access(path, os.R_OK): if href and os.access(path, os.R_OK):
return open(path, 'rb').read() return open(path, 'rb').read()
def extract_calibre_cover(raw, base, log):
from calibre.ebooks.BeautifulSoup import BeautifulSoup
soup = BeautifulSoup(raw)
matches = soup.find(name=['h1', 'h2', 'h3', 'h4', 'h5', 'h6', 'p', 'span',
'font', 'br'])
images = soup.findAll('img')
if matches is None and len(images) == 1 and \
images[0].get('alt', '')=='cover':
img = images[0]
img = os.path.join(base, *img['src'].split('/'))
if os.path.exists(img):
return open(img, 'rb').read()
def render_html_svg_workaround(path_to_html, log, width=590, height=750): def render_html_svg_workaround(path_to_html, log, width=590, height=750):
from calibre.ebooks.oeb.base import SVG_NS from calibre.ebooks.oeb.base import SVG_NS
raw = open(path_to_html, 'rb').read() raw = open(path_to_html, 'rb').read()
@ -80,6 +93,11 @@ def render_html_svg_workaround(path_to_html, log, width=590, height=750):
os.path.dirname(path_to_html), log) os.path.dirname(path_to_html), log)
except: except:
pass pass
if data is None:
try:
data = extract_calibre_cover(raw, os.path.dirname(path_to_html), log)
except:
pass
if data is None: if data is None:
renderer = render_html(path_to_html, width, height) renderer = render_html(path_to_html, width, height)
data = getattr(renderer, 'data', None) data = getattr(renderer, 'data', None)