From e1f7c784ad83e0252b682f780d852ecc1c2f1eb5 Mon Sep 17 00:00:00 2001 From: j Date: Mon, 5 Nov 2018 21:18:59 +0100 Subject: [PATCH] use pdftocairo for pdf preview --- pandora/document/utils.py | 17 ++++++++++++----- 1 file changed, 12 insertions(+), 5 deletions(-) diff --git a/pandora/document/utils.py b/pandora/document/utils.py index ee28ff53..3df53197 100644 --- a/pandora/document/utils.py +++ b/pandora/document/utils.py @@ -1,4 +1,5 @@ # -*- coding: utf-8 -*- +import os import re import subprocess @@ -22,11 +23,17 @@ def pdfinfo(pdf): return data def extract_pdfpage(pdf, image, page): - page -= 1 - cmd = ['convert', '%s[%d]' % (pdf, page), - '-background', 'white', '-flatten', '-resize', '1024x1024', image] - p = subprocess.Popen(cmd, close_fds=True) - p.wait() + page = str(page) + cmd = [ + 'pdftocairo', + '-jpeg', + '-f', page, '-l', page, + '-singlefile', + '-scale-to', '2048', + pdf, + os.path.splitext(image)[0] + ] + subprocess.call(cmd, stdout=open('/dev/null', 'wb')) return image def get_documents(text):