sporada/ocr
1
0
Fork 0
mirror of https://gitlab.ub.uni-bielefeld.de/sfb1288inf/ocr.git synced 2025-10-31 19:53:16 +00:00
Code Issues Projects Releases Wiki Activity

Change output files file format

Browse Source
This commit is contained in:
Patrick Jentsch
2022-01-14 10:56:16 +01:00
parent c640d9743f
commit f51a8c4546
1 changed files with 12 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files

22
ocr
View File

@@ -488,39 +488,41 @@ class MainWorkflow(WorkflowRunner):
# Remove temporary directory
os.rmdir(job.tmp_dir)
# Track output files
relative_input = os.path.relpath(job.file, start=self.input_dir)
relative_output_dir = os.path.relpath(job.output_dir, start=self.output_dir) # noqa
for x in os.listdir(os.path.join(job.output_dir, 'images')):
self.output_files.append(
{
'directory': os.path.join(os.path.relpath(job.output_dir, start=self.output_dir), 'images'), # noqa
'filename': x,
'input': relative_input,
'path': os.path.join(relative_output_dir, 'images', x),
'mimetype': 'image/png'
}
)
self.output_files.append(
{
'directory': os.path.relpath(job.output_dir, start=self.output_dir), # noqa
'filename': '{}.hocr'.format(job.name),
'input': relative_input,
'path': os.path.join(relative_output_dir, '{}.hocr'.format(job.name)), # noqa
'mimetype': 'application/xhtml+xml'
}
)
self.output_files.append(
{
'directory': os.path.relpath(job.output_dir, start=self.output_dir), # noqa
'filename': '{}.pdf'.format(job.name),
'input': relative_input,
'filename': os.path.join(relative_output_dir, '{}.pdf'.format(job.name)), # noqa
'mimetype': 'application/pdf'
}
)
self.output_files.append(
{
'directory': os.path.relpath(job.output_dir, start=self.output_dir), # noqa
'filename': '{}.txt'.format(job.name),
'input': relative_input,
'filename': os.path.join(relative_output_dir, '{}.txt'.format(job.name)), # noqa
'mimetype': 'text/plain'
}
)
self.output_files.append(
{
'directory': os.path.relpath(job.output_dir, start=self.output_dir), # noqa
'filename': '{}.xml'.format(job.name),
'input': relative_input,
'filename': os.path.join(relative_output_dir, '{}.xml'.format(job.name)), # noqa
'mimetype': 'application/tei+xml'
}
)