sporada/ocr
1
0
Fork 0
mirror of https://gitlab.ub.uni-bielefeld.de/sfb1288inf/ocr.git synced 2024-12-26 17:34:18 +00:00
Code Issues Projects Releases Wiki Activity

Add zip fielname argument

Browse Source
This commit is contained in:
stephan 2020-02-17 14:26:50 +01:00
parent 6c4a642cb7
commit 880f0efcf9
1 changed files with 14 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

24
ocr
View File

@ -67,9 +67,9 @@ def parse_arguments():
) )
parser.add_argument( parser.add_argument(
'--zip', '--zip',
action='store_true', default='ocr-result-files',
default=False,
dest='zip', dest='zip',
type=str,
help='package result files in zip bundles', help='package result files in zip bundles',
required=False required=False
) )
@ -400,8 +400,9 @@ class OCRWorkflow(WorkflowRunner):
all_zip_job_dependencies = (hocr_to_tei_jobs all_zip_job_dependencies = (hocr_to_tei_jobs
+ pdf_merge_jobs + pdf_merge_jobs
+ txt_merge_jobs) + txt_merge_jobs)
cmd = 'cd "%s" && zip all.zip */*.{pdf,txt,xml} -x "pyflow.data*" && cd -' % ( cmd = 'cd "%s" && zip "%s"-all-files.zip */*.{pdf,txt,xml} -x "pyflow.data*" && cd -' % (
self.output_dir self.output_dir,
self.zip
) )
all_zip_jobs.append( all_zip_jobs.append(
self.addTask( self.addTask(
@ -413,8 +414,9 @@ class OCRWorkflow(WorkflowRunner):
pdf_zip_jobs = [] pdf_zip_jobs = []
pdf_zip_job_dependencies = all_zip_jobs pdf_zip_job_dependencies = all_zip_jobs
cmd = 'cd "%s" && zip -m pdf.zip */*.pdf -x "pyflow.data*" && cd -' % ( cmd = 'cd "%s" && zip -m "%s"-pdf.zip */*.pdf -x "pyflow.data*" && cd -' % (
self.output_dir self.output_dir,
self.zip
) )
pdf_zip_jobs.append( pdf_zip_jobs.append(
self.addTask( self.addTask(
@ -426,8 +428,9 @@ class OCRWorkflow(WorkflowRunner):
txt_zip_jobs = [] txt_zip_jobs = []
txt_zip_job_dependencies = all_zip_jobs txt_zip_job_dependencies = all_zip_jobs
cmd = 'cd "%s" && zip -m txt.zip */*.txt -x "pyflow.data*" && cd -' % ( cmd = 'cd "%s" && zip -m "%s"-txt.zip */*.txt -x "pyflow.data*" && cd -' % (
self.output_dir self.output_dir,
self.zip
) )
txt_zip_jobs.append( txt_zip_jobs.append(
self.addTask( self.addTask(
@ -439,8 +442,9 @@ class OCRWorkflow(WorkflowRunner):
xml_zip_jobs = [] xml_zip_jobs = []
xml_zip_job_dependencies = all_zip_jobs xml_zip_job_dependencies = all_zip_jobs
cmd = 'cd "%s" && zip -m xml.zip */*.xml -x "pyflow.data*" && cd -' % ( cmd = 'cd "%s" && zip -m "%s"-xml.zip */*.xml -x "pyflow.data*" && cd -' % (
self.output_dir self.output_dir,
self.zip
) )
xml_zip_jobs.append( xml_zip_jobs.append(
self.addTask( self.addTask(