0x2620/openmedialibrary
1
0
Fork 0
Code Issues 131 Pull requests Releases Wiki Activity

fix epub metadata parser

Browse source
This commit is contained in:
j 2015-12-01 17:20:32 +01:00
parent 20864c591f
commit 81cd9c2337
2 changed files with 6 additions and 4 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
oml/media/__init__.py
View file

@ -48,7 +48,7 @@ def metadata(f, from_=None):
): ):
if key in info: if key in info:
value = info[key] value = info[key]
if isinstance(value, str): if isinstance(value, bytes):
try: try:
value = value.decode('utf-8') value = value.decode('utf-8')
except: except:

8
oml/media/epub.py
View file

@ -77,18 +77,20 @@ def info(epub):
info = ET.fromstring(z.read(opf[0])) info = ET.fromstring(z.read(opf[0]))
metadata = info.findall('{http://www.idpf.org/2007/opf}metadata')[0] metadata = info.findall('{http://www.idpf.org/2007/opf}metadata')[0]
for e in metadata.getchildren(): for e in metadata.getchildren():
if e.text and e.text not in ('unknown', 'none'): if e.text and e.text.strip() and e.text not in ('unknown', 'none'):
key = e.tag.split('}')[-1] key = e.tag.split('}')[-1]
key = { key = {
'creator': 'author', 'creator': 'author',
}.get(key, key) }.get(key, key)
value = e.text value = e.text.strip()
if key == 'identifier': if key == 'identifier':
value = normalize_isbn(value) value = normalize_isbn(value)
if stdnum.isbn.is_valid(value): if stdnum.isbn.is_valid(value):
data['isbn'] = [value] data['isbn'] = [value]
elif key == 'author':
data[key] = value.split(', ')
else: else:
data[key] = e.text data[key] = value
text = extract_text(epub) text = extract_text(epub)
data['textsize'] = len(text) data['textsize'] = len(text)
if not 'isbn' in data: if not 'isbn' in data: