From 265432f2a57135fd47c40e008de84a858b370e62 Mon Sep 17 00:00:00 2001 From: jonaswinkler <17569239+jonaswinkler@users.noreply.github.com> Date: Sun, 21 Feb 2021 23:39:19 +0100 Subject: [PATCH] fix up the ocrmypdf parameter construction for clean-final and redo --- src/paperless_tesseract/parsers.py | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/src/paperless_tesseract/parsers.py b/src/paperless_tesseract/parsers.py index 728c40a56..deadf2234 100644 --- a/src/paperless_tesseract/parsers.py +++ b/src/paperless_tesseract/parsers.py @@ -155,8 +155,11 @@ class RasterisedDocumentParser(DocumentParser): if settings.OCR_CLEAN == 'clean': ocrmypdf_args['clean'] = True - elif settings.OCR_CLEAN == 'clean-final' and not settings.OCR_MODE == 'redo': # NOQA: E501 - ocrmypdf_args['clean_final'] = True + elif settings.OCR_CLEAN == 'clean-final': + if settings.OCR_MODE == 'redo': + ocrmypdf_args['clean'] = True + else: + ocrmypdf_args['clean_final'] = True if settings.OCR_DESKEW and not settings.OCR_MODE == 'redo': ocrmypdf_args['deskew'] = True