From 3b53e9adac014aa46ab7a2f378632feef5271755 Mon Sep 17 00:00:00 2001 From: "James R. Barlow" Date: Mon, 11 Jan 2016 17:22:50 -0800 Subject: [PATCH] Use tesseract cache for -psm --- tests/spoof/tesseract_cache.py | 5 +++++ tests/test_main.py | 4 ++-- 2 files changed, 7 insertions(+), 2 deletions(-) diff --git a/tests/spoof/tesseract_cache.py b/tests/spoof/tesseract_cache.py index df83a67f..f6586808 100755 --- a/tests/spoof/tesseract_cache.py +++ b/tests/spoof/tesseract_cache.py @@ -37,6 +37,11 @@ def main(): m.update(lang.encode()) except ValueError: pass + try: + psm = sys.argv[sys.argv.index('-psm') + 1] + m.update(psm.encode()) + except ValueError: + pass input_file = sys.argv[-3] output_file = sys.argv[-2] diff --git a/tests/test_main.py b/tests/test_main.py index c011ca6b..3a652dfc 100644 --- a/tests/test_main.py +++ b/tests/test_main.py @@ -354,12 +354,12 @@ def test_encrypted(): 'hocr', 'tesseract', ]) -def test_pagesegmode(renderer): +def test_pagesegmode(renderer, spoof_tesseract_cache): check_ocrmypdf( 'skew.pdf', 'test_psm_%s.pdf' % renderer, '--tesseract-pagesegmode', '7', '-v', '1', - '--pdf-renderer', renderer) + '--pdf-renderer', renderer, env=spoof_tesseract_cache)