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

port to python3

Browse source
This commit is contained in:
j 2014-09-03 00:32:44 +02:00
commit 8e27b9f76e
51 changed files with 272 additions and 248 deletions
Download patch file Download diff file Expand all files Collapse all files

10
oml/media/pdf.py
View file

@ -1,6 +1,6 @@
# -*- coding: utf-8 -*-
# vi:si:et:sw=4:sts=4:ts=4
from __future__ import division
import sys
import tempfile
@ -9,7 +9,7 @@ import os
import shutil
from glob import glob
from pyPdf import PdfFileReader
#from pyPdf import PdfFileReader
import stdnum.isbn
import settings
@ -139,9 +139,9 @@ def info(pdf):
if stdnum.isbn.is_valid(value):
data['isbn'] = [value]
del data['identifier']
for key, value in data.iteritems():
for key, value in data.items():
if isinstance(value, dict):
value = ' '.join(value.values())
value = ' '.join(list(value.values()))
data[key] = value
text = extract_text(pdf)
data['textsize'] = len(text)
@ -150,7 +150,7 @@ def info(pdf):
isbn = extract_isbn(text)
if isbn:
data['isbn'] = [isbn]
if 'isbn' in data and isinstance(data['isbn'], basestring):
if 'isbn' in data and isinstance(data['isbn'], str):
data['isbn'] = [data['isbn']]
if 'date' in data and len(data['date']) == 8 and data['date'].isdigit():
d = data['date']