From 7313a77c2a0f8d0e20b61a50b55a13fd18420498 Mon Sep 17 00:00:00 2001 From: "James R. Barlow" Date: Thu, 17 Dec 2015 11:41:54 -0800 Subject: [PATCH] Implement pdf renderer side of tess spoof --- .../spoof/{tesseract_hocr_empty.py => tesseract_noop.py} | 7 +++++++ tests/test_main.py | 8 ++++---- 2 files changed, 11 insertions(+), 4 deletions(-) rename tests/spoof/{tesseract_hocr_empty.py => tesseract_noop.py} (87%) diff --git a/tests/spoof/tesseract_hocr_empty.py b/tests/spoof/tesseract_noop.py similarity index 87% rename from tests/spoof/tesseract_hocr_empty.py rename to tests/spoof/tesseract_noop.py index 70a15a8c..ee7cadae 100755 --- a/tests/spoof/tesseract_hocr_empty.py +++ b/tests/spoof/tesseract_noop.py @@ -1,5 +1,6 @@ #!/usr/bin/env python3 import sys +import img2pdf VERSION_STRING = '''tesseract 3.04.00 @@ -42,6 +43,12 @@ def main(): output = sys.argv[-2] with open(output + '.hocr', 'w', encoding='utf-8') as f: f.write(HOCR_TEMPLATE.format('1000', '1000')) + elif sys.argv[-1] == 'pdf': + inputf = sys.argv[-3] + output = sys.argv[-2] + pdf_bytes = img2pdf.convert([inputf], dpi=300) + with open(output + '.pdf', 'wb') as f: + f.write(pdf_bytes) else: print("Spoof doesn't understand arguments", file=sys.stderr) print(sys.argv, file=sys.stderr) diff --git a/tests/test_main.py b/tests/test_main.py index 1240d842..9c12a1f0 100644 --- a/tests/test_main.py +++ b/tests/test_main.py @@ -83,18 +83,18 @@ def test_quick(): @pytest.fixture -def spoof_tesseract_hocr_empty(): +def spoof_tesseract_noop(): env = os.environ.copy() - program = os.path.join(SPOOF_PATH, 'tesseract_hocr_empty.py') + program = os.path.join(SPOOF_PATH, 'tesseract_noop.py') check_call(['chmod', "+x", program]) env['OCRMYPDF_TESSERACT'] = program return env -def test_deskew(spoof_tesseract_hocr_empty): +def test_deskew(spoof_tesseract_noop): # Run with deskew deskewed_pdf = check_ocrmypdf( - 'skew.pdf', 'test_deskew.pdf', '-d', env=spoof_tesseract_hocr_empty) + 'skew.pdf', 'test_deskew.pdf', '-d', env=spoof_tesseract_noop) # Now render as an image again and use Leptonica to find the skew angle # to confirm that it was deskewed