sporada/ocr
1
0
Fork 0
mirror of https://gitlab.ub.uni-bielefeld.de/sfb1288inf/ocr.git synced 2024-12-26 18:24:17 +00:00
Code Issues Projects Releases Wiki Activity

Remove unused dependencies and use ghostscript for image split

Browse Source
This commit is contained in:
Patrick Jentsch 2020-07-01 11:03:34 +02:00
parent aee9628e5e
commit 2b63ba9e59
2 changed files with 24 additions and 16 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
Dockerfile
View File

@ -14,7 +14,6 @@ RUN apt-get update \
ca-certificates \ ca-certificates \
gnupg2 \ gnupg2 \
ghostscript \ ghostscript \
imagemagick \
poppler-utils \ poppler-utils \
python2.7 \ python2.7 \
python3.7 \ python3.7 \

39
ocr
View File

@ -98,12 +98,16 @@ class OCRPipeline(WorkflowRunner):
n_cores = min(self.n_cores, max(1, int(self.n_cores / len(self.jobs)))) n_cores = min(self.n_cores, max(1, int(self.n_cores / len(self.jobs))))
for i, job in enumerate(self.jobs): for i, job in enumerate(self.jobs):
output_dir = os.path.join(job.output_dir, 'tmp') output_dir = os.path.join(job.output_dir, 'tmp')
output_file_base = os.path.join(output_dir, 'page') cmd = 'gs'
cmd = 'pdftoppm' cmd += ' -dBATCH'
cmd += ' -r 300' cmd += ' -dNOPAUSE'
cmd += ' -tiff' cmd += ' -dNumRenderingThreads={}'.format(n_cores)
cmd += ' -tiffcompression lzw' cmd += ' -dQUIET'
cmd += ' "{}" "{}"'.format(job.file, output_file_base) cmd += ' -r300'
cmd += ' -sDEVICE=tiff24nc'
cmd += ' -sCompression=lzw'
cmd += ' "-sOutputFile={}/page-%d.tif"'.format(output_dir)
cmd += ' "{}"'.format(job.file)
deps = 'setup_output_directory_-_{}'.format(i) deps = 'setup_output_directory_-_{}'.format(i)
lbl = 'split_input_-_{}'.format(i) lbl = 'split_input_-_{}'.format(i)
split_input_jobs.append(self.addTask(command=cmd, split_input_jobs.append(self.addTask(command=cmd,
@ -138,8 +142,9 @@ class OCRPipeline(WorkflowRunner):
files.sort(key=lambda x: int(re.search(r'\d+', x).group(0))) files.sort(key=lambda x: int(re.search(r'\d+', x).group(0)))
files = map(lambda x: os.path.join(input_dir, x), files) files = map(lambda x: os.path.join(input_dir, x), files)
cmd = 'ocropus-nlbin "{}"'.format('" "'.join(files)) cmd = 'ocropus-nlbin "{}"'.format('" "'.join(files))
cmd += ' -o "{}"'.format(output_dir) cmd += ' --nocheck'
cmd += ' -Q "{}"'.format(n_cores) cmd += ' --output "{}"'.format(output_dir)
cmd += ' --parallel "{}"'.format(n_cores)
deps = 'split_input_-_{}'.format(i) deps = 'split_input_-_{}'.format(i)
lbl = 'binarization_-_{}'.format(i) lbl = 'binarization_-_{}'.format(i)
binarization_jobs.append(self.addTask(command=cmd, binarization_jobs.append(self.addTask(command=cmd,
@ -322,19 +327,22 @@ class OCRPipeline(WorkflowRunner):
' ################################################## ' ##################################################
''' '''
pdf_compression_jobs = [] pdf_compression_jobs = []
n_cores = min(self.n_cores, max(1, int(self.n_cores / len(self.jobs))))
if self.compress: if self.compress:
for i, job in enumerate(self.jobs): for i, job in enumerate(self.jobs):
file = filter(lambda x: x.endswith('.pdf'), os.listdir(job.output_dir))[0] # noqa file = filter(lambda x: x.endswith('.pdf'), os.listdir(job.output_dir))[0] # noqa
original_file = os.path.join(job.output_dir, file) original_file = os.path.join(job.output_dir, file)
compressed_file = os.path.join(job.output_dir, 'c_' + file) compressed_file = os.path.join(job.output_dir, 'c_' + file)
cmd = 'gs' cmd = 'gs'
cmd += ' -sDEVICE=pdfwrite'
cmd += ' -dCompatibilityLevel=1.4'
cmd += ' -dPDFSETTINGS=/ebook'
cmd += ' -dNOPAUSE'
cmd += ' -dQUIET'
cmd += ' -dBATCH' cmd += ' -dBATCH'
cmd += ' -sOutputFile="{}"'.format(compressed_file) cmd += ' -dNOPAUSE'
cmd += ' -dNumRenderingThreads={}'.format(n_cores)
cmd += ' -dPDFSETTINGS=/ebook'
# -dCompatibilityLevel must be defined after -dPDFSETTINGS
cmd += ' -dCompatibilityLevel=1.4'
cmd += ' -dQUIET'
cmd += ' -sDEVICE=pdfwrite'
cmd += ' "-sOutputFile={}"'.format(compressed_file)
cmd += ' "{}"'.format(original_file) cmd += ' "{}"'.format(original_file)
cmd += ' && ' cmd += ' && '
cmd += 'mv "{}" "{}"'.format(compressed_file, original_file) cmd += 'mv "{}" "{}"'.format(compressed_file, original_file)
@ -342,7 +350,8 @@ class OCRPipeline(WorkflowRunner):
lbl = 'pdf_compression_-_{}'.format(i) lbl = 'pdf_compression_-_{}'.format(i)
pdf_compression_jobs.append(self.addTask(command=cmd, pdf_compression_jobs.append(self.addTask(command=cmd,
dependencies=deps, dependencies=deps,
label=lbl)) label=lbl,
nCores=n_cores))
''' '''
' ################################################## ' ##################################################