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

Use PoDoFo in pdfmanipulate info

Browse Source
This commit is contained in:
Kovid Goyal 2009-05-31 09:34:27 -07:00
parent c3db153421
commit a52242cecb
2 changed files with 63 additions and 31 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

56
src/calibre/ebooks/pdf/manipulate/info.py
View File

@ -9,16 +9,14 @@ __docformat__ = 'restructuredtext en'
Merge PDF files into a single PDF document.
'''
import os, re, sys, time
from optparse import OptionGroup, Option
import os, sys
from calibre.utils.config import OptionParser
from calibre.utils.logging import Log
from calibre.constants import preferred_encoding
from calibre.customize.conversion import OptionRecommendation
from calibre.ebooks.pdf.verify import is_valid_pdfs, is_encrypted, is_encrypted
from pyPdf import PdfFileWriter, PdfFileReader
from calibre.ebooks.pdf.verify import is_valid_pdfs, is_encrypted
from calibre import prints
from calibre.utils.podofo import podofo, podofo_err
USAGE = '\n%prog %%name ' + _('''\
file.pdf ...
@ -35,40 +33,36 @@ def option_parser(name):
return OptionParser(usage=usage)
def print_info(pdf_path):
with open(os.path.abspath(pdf_path), 'rb') as pdf_file:
pdf = PdfFileReader(pdf_file)
print _('Title: %s' % pdf.documentInfo.title)
print _('Author: %s' % pdf.documentInfo.author)
print _('Subject: %s' % pdf.documentInfo.subject)
print _('Creator: %s' % pdf.documentInfo.creator)
print _('Producer: %s' % pdf.documentInfo.producer)
#print _('Creation Date: %s' % time.strftime('%a %b %d %H:%M:%S %Y', time.gmtime(os.path.getctime(pdf_path))))
#print _('Modification Date: %s' % time.strftime('%a %b %d %H:%M:%S %Y', time.gmtime(os.path.getmtime(pdf_path))))
print _('Pages: %s' % pdf.numPages)
#print _('Encrypted: %s' % pdf.isEncrypted)
try:
print _('File Size: %s bytes' % os.path.getsize(pdf_path))
except: pass
try:
pdf_file.seek(0)
vline = pdf_file.readline()
mo = re.search('(?iu)^%...-(?P<version>\d+\.\d+)', vline)
if mo != None:
print _('PDF Version: %s' % mo.group('version'))
except: pass
if not podofo:
raise RuntimeError('Failed to load PoDoFo with error:'+podofo_err)
p = podofo.PDFDoc()
p.open(pdf_path)
fmt = lambda x, y: '%-20s: %s'%(x, y)
print
prints(fmt(_('Title'), p.title))
prints(fmt(_('Author'), p.author))
prints(fmt(_('Subject'), p.subject))
prints(fmt(_('Creator'), p.creator))
prints(fmt(_('Producer'), p.producer))
prints(fmt(_('Pages'), p.pages))
prints(fmt(_('File Size'), os.stat(pdf_path).st_size))
prints(fmt(_('PDF Version'), p.version if p.version else _('Unknown')))
def main(args=sys.argv, name=''):
log = Log()
parser = option_parser(name)
opts, args = parser.parse_args(args)
args = args[1:]
if len(args) < 1:
print 'Error: No PDF sepecified.\n'
print_help(parser, log)
return 1
bad_pdfs = is_valid_pdfs(args)
if bad_pdfs != []:
for pdf in bad_pdfs:
@ -85,7 +79,7 @@ def main(args=sys.argv, name=''):
for pdf in args:
print_info(pdf)
return 0
if __name__ == '__main__':

38
src/calibre/utils/podofo/podofo.cpp
View File

@ -108,6 +108,40 @@ podofo_PDFDoc_pages_getter(podofo_PDFDoc *self, void *closure) {
return ans;
}
static PyObject *
podofo_PDFDoc_version_getter(podofo_PDFDoc *self, void *closure) {
int version;
try {
version = self->doc->GetPdfVersion();
} catch(const PdfError & err) {
podofo_set_exception(err);
return NULL;
}
switch(version) {
case ePdfVersion_1_0:
return Py_BuildValue("s", "1.0");
case ePdfVersion_1_1:
return Py_BuildValue("s", "1.1");
case ePdfVersion_1_2:
return Py_BuildValue("s", "1.2");
case ePdfVersion_1_3:
return Py_BuildValue("s", "1.3");
case ePdfVersion_1_4:
return Py_BuildValue("s", "1.4");
case ePdfVersion_1_5:
return Py_BuildValue("s", "1.5");
case ePdfVersion_1_6:
return Py_BuildValue("s", "1.6");
case ePdfVersion_1_7:
return Py_BuildValue("s", "1.7");
default:
return Py_BuildValue("");
}
return Py_BuildValue("");
}
static PyObject *
podofo_PDFDoc_delete_pages(podofo_PDFDoc *self, PyObject *args, PyObject *kwargs) {
int first_page, num_pages;
@ -315,6 +349,10 @@ static PyGetSetDef podofo_PDFDoc_getsetters[] = {
(getter)podofo_PDFDoc_pages_getter, NULL,
(char *)"Number of pages in document (read only)",
NULL},
{(char *)"version",
(getter)podofo_PDFDoc_version_getter, NULL,
(char *)"The PDF version (read only)",
NULL},
{NULL} /* Sentinel */
};