Fix issues with —pdf-renderer tess4 page skipping

If tess4 renderer needed to skip OCR on a page it would end up
duplicating the page contents onto the new page, rather than creating
a blank OCR layer and placing it on the output page. This created
duplicated content in output files.

ocrmypdf/exec/tesseract.py

@@ -11,6 +11,7 @@ from ..helpers import page_number
 from . import get_program
 from collections import namedtuple
 from textwrap import dedent
+import PyPDF2 as pypdf
 
 from subprocess import Popen, PIPE, CalledProcessError, \
     TimeoutExpired, check_output, STDOUT, DEVNULL
@@ -272,6 +273,24 @@ def generate_hocr(input_file, output_hocr, language: list, engine_mode,
                 f_out.write(line)
 
 
+def use_skip_page(text_only, skip_pdf, output_pdf):
+    if not text_only:
+        os.symlink(skip_pdf, output_pdf)
+        return
+
+    # For text only we must create a blank page with dimensions identical
+    # to the skip page because this is equivalent to a page with no text
+
+    pdf_in = pypdf.PdfFileReader(skip_pdf)
+    page0 = pdf_in.pages[0]
+
+    with open(output_pdf, 'wb') as out:
+        pdf_out = pypdf.PdfFileWriter()
+        w, h = page0.mediaBox.getWidth(), page0.mediaBox.getHeight()
+        pdf_out.addBlankPage(w, h)
+        pdf_out.write(out)
+
+
 def generate_pdf(input_image, skip_pdf, output_pdf, language: list,
                  engine_mode, text_only: bool,
                  tessconfig: list, timeout: float, pagesegmode: int, log):
@@ -309,14 +328,14 @@ def generate_pdf(input_image, skip_pdf, output_pdf, language: list,
             universal_newlines=True, timeout=timeout)
     except TimeoutExpired:
         page_timedout(log, input_image)
-        shutil.copy(skip_pdf, output_pdf)
+        use_skip_page(text_only, skip_pdf, output_pdf)
     except CalledProcessError as e:
         tesseract_log_output(log, e.output, input_image)
         if 'read_params_file: parameter not found' in e.output:
             raise TesseractConfigError() from e
 
         if 'Image too large' in e.output:
-            shutil.copy(skip_pdf, output_pdf)
+            use_skip_page(text_only, skip_pdf, output_pdf)
             return
         raise e from e
     else:

ocrmypdf/pipeline.py

@@ -773,7 +773,8 @@ def generate_postscript_stub(
 def skip_page(
         input_file,
         output_file,
-        log):
+        log,
+        context):
     # The purpose of this step is its filter to forward only the skipped
     # files (.skip.oriented.pdf) while disregarding the processed ones
     # (.ocr.oriented.pdf).  Alternative would be for merge_pages to filter
@@ -1048,7 +1049,7 @@ def build_pipeline(options, work_folder, log, context):
         filter=suffix('.skip.oriented.pdf'),
         output='.done.pdf',
         output_dir=work_folder,
-        extras=[log])
+        extras=[log, context])
 
     # Merge pages
     task_merge_pages_ghostscript = main_pipeline.merge(

tests/test_tess4.py

@@ -6,11 +6,42 @@ from ocrmypdf.exceptions import ExitCode
 from ocrmypdf.exec import tesseract
 from ocrmypdf import pageinfo
 import sys
+import os
+import PyPDF2 as pypdf
+
+
+spoof = pytest.helpers.spoof
+
+
+def tess4_possible_location():
+    """The location of tesseract 4 may be OCRMYPDF_TESS4, OCRMYPDF_TESSERACT,
+    or the installed version on PATH."""
+    return os.environ.get('OCRMYPDF_TESS4') or \
+           os.environ.get('OCRMYPDF_TESSERACT') or \
+            'tesseract'
+
+
+@pytest.fixture
+def ensure_tess4():
+    return spoof(tesseract=tess4_possible_location())
+
+
+def tess4_available():
+    """Check if a tesseract 4 binary is available, even if it's not the
+    official "tesseract" on PATH
+
+    """
+    old_environ = os.environ.copy()
+    try:
+        os.environ['OCRMYPDF_TESSERACT'] = tess4_possible_location()
+        return tesseract.v4() and tesseract.has_textonly_pdf()
+    finally:
+        os.environ = old_environ
 
 
 # Skip all tests in this file if not tesseract 4
 pytestmark = pytest.mark.skipif(
-    not (tesseract.v4() and tesseract.has_textonly_pdf()),
+    not tess4_available(),
     reason="tesseract 4.0 with textonly_pdf feature required")
 
 check_ocrmypdf = pytest.helpers.check_ocrmypdf
@@ -18,14 +49,15 @@ run_ocrmypdf = pytest.helpers.run_ocrmypdf
 spoof = pytest.helpers.spoof
 
 
-def test_textonly_pdf(resources, outdir):
+def test_textonly_pdf(ensure_tess4, resources, outdir):
     check_ocrmypdf(
         resources / 'linn.pdf',
-        outdir / 'linn_textonly.pdf', '--pdf-renderer', 'tess4')
+        outdir / 'linn_textonly.pdf', '--pdf-renderer', 'tess4',
+        env=ensure_tess4)
 
 
 @pytest.mark.skipif(sys.version_info < (3, 5), reason="needs math.isclose")
-def test_pagesize_consistency_tess4(resources, outpdf):
+def test_pagesize_consistency_tess4(ensure_tess4, resources, outpdf):
     from math import isclose
 
     infile = resources / 'linn.pdf'
@@ -35,9 +67,33 @@ def test_pagesize_consistency_tess4(resources, outpdf):
     check_ocrmypdf(
         infile,
         outpdf, '--pdf-renderer', 'tess4',
-        '--clean', '--deskew', '--remove-background', '--clean-final')
+        '--clean', '--deskew', '--remove-background', '--clean-final',
+        env=ensure_tess4)
 
     after_dims = pytest.helpers.first_page_dimensions(outpdf)
 
     assert isclose(before_dims[0], after_dims[0])
     assert isclose(before_dims[1], after_dims[1])
+
+
+@pytest.mark.parametrize('basename', ['graph_ocred.pdf', 'cardinal.pdf'])
+def test_skip_pages_does_not_replicate(
+        ensure_tess4, resources, basename, outdir):
+    infile = resources / basename
+    outpdf = outdir / basename
+
+    check_ocrmypdf(
+        infile,
+        outpdf, '--pdf-renderer', 'tess4', '--force-ocr',
+        '--tesseract-timeout', '0',
+        env=ensure_tess4
+    )
+
+    info_in = pageinfo.pdf_get_all_pageinfo(str(infile))
+
+    info = pageinfo.pdf_get_all_pageinfo(str(outpdf))
+    for page in info:
+        assert len(page['images']) == 1, "skipped page was replicated"
+
+    for n in range(len(info_in)):
+        assert info[n]['width_inches'] == info_in[n]['width_inches']