pestctrl/guess-filename.py
1
0
Fork 0
forked from Github/guess-filename.py
Code Issues Pull requests Projects Releases Packages Wiki Activity

move check for PDF file before loading PDF parsing library

Browse source
This commit is contained in:
Karl Voit 2017-12-08 15:44:05 +01:00
parent 0a50af236a
commit 4495f83aff
1 changed files with 7 additions and 7 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

14
guessfilename.py
View file

@ -1,6 +1,6 @@
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
PROG_VERSION = u"Time-stamp: <2017-12-08 15:21:44 vk>"
PROG_VERSION = u"Time-stamp: <2017-12-08 15:40:50 vk>"
# TODO:
@ -627,6 +627,12 @@ class GuessFilename(object):
filename = os.path.join(dirname, basename)
assert os.path.isfile(filename)
datetimestr, basefilename, tags, extension = self.split_filename_entities(basename)
if extension.lower() != 'pdf':
logging.debug("File is not a PDF file and thus can't be parsed by this script: %s" % filename)
return False
try:
pdffile = PyPDF2.PdfFileReader(open(filename, "rb"))
except:
@ -645,12 +651,6 @@ class GuessFilename(object):
logging.warning('Could read PDF file content but it is empty (skipping content analysis)')
return False
datetimestr, basefilename, tags, extension = self.split_filename_entities(basename)
if extension.lower() != 'pdf':
logging.debug("File is not a PDF file and thus can't be parsed by this script: %s" % filename)
return False
# 2010-06-08 easybank - neue TAN-Liste -- scan private.pdf
if self.fuzzy_contains_all_of(content, ["Transaktionsnummern (TANs)", "Ihre TAN-Liste in Verlust geraten"]) and \
datetimestr: