mirror of
https://github.com/qurator-spk/ocrd-galley.git
synced 2025-06-09 14:49:53 +02:00
✨ Use MAX file group name instead of BEST
Some checks failed
continuous-integration/drone/push Build is failing
Some checks failed
continuous-integration/drone/push Build is failing
We were using the file group name BEST for what Kitodo seems to call MAX by convention. So we use MAX now. Currently, we work under the assumption that, if MAX exists in the METS retrieved by OAI-PMH, it's not what we want and we replace it with our own IIIF URLS with full size. Fixes GH-43.
This commit is contained in:
parent
94a035e2cf
commit
691be243f6
2 changed files with 10 additions and 5 deletions
13
ppn2ocr
13
ppn2ocr
|
@ -91,15 +91,20 @@ def make_workspace(ppn, workspace):
|
|||
remove_file_grp(mets, 'PRESENTATION')
|
||||
remove_file_grp(mets, 'LOCAL')
|
||||
|
||||
# Duplicate DEFAULT file group into a new file group BEST
|
||||
|
||||
# Delete MAX file group - we assume that, if it exists, it is not as
|
||||
# we expect it, e.g. IIIF full URLs
|
||||
remove_file_grp(mets, 'MAX')
|
||||
|
||||
# Duplicate DEFAULT file group into a new file group MAX
|
||||
format_ = 'tif'
|
||||
file_grp_default = mets.find('//mets:fileGrp[@USE="DEFAULT"]', namespaces=XMLNS)
|
||||
file_grp_best = deepcopy(file_grp_default)
|
||||
|
||||
file_grp_best.attrib['USE'] = 'BEST'
|
||||
file_grp_best.attrib['USE'] = 'MAX'
|
||||
for f in file_grp_best.findall('./mets:file', namespaces=XMLNS):
|
||||
old_id = f.attrib['ID']
|
||||
new_id = re.sub('DEFAULT', 'BEST', old_id)
|
||||
new_id = re.sub('DEFAULT', 'MAX', old_id)
|
||||
f.attrib['ID'] = new_id
|
||||
f.attrib['MIMETYPE'] = mime_type_for_format(format_)
|
||||
|
||||
|
@ -157,7 +162,7 @@ def ppn2ocr(ppn):
|
|||
# XXX
|
||||
# subprocess.run([
|
||||
# os.path.join(self_dir, 'run-docker-hub'),
|
||||
# '-I', 'BEST',
|
||||
# '-I', 'MAX',
|
||||
# '--skip-validation'
|
||||
# ])
|
||||
|
||||
|
|
Loading…
Add table
Add a link
Reference in a new issue