qurator-spk/ocrd-galley
1
0
Fork 0
mirror of https://github.com/qurator-spk/ocrd-galley.git synced 2025-07-04 02:39:51 +02:00
Code Issues Projects Releases Wiki Activity

🚧 ppn2ocr: Properly remove the PRESENTATION file group

Browse source
This commit is contained in:
Gerber, Mike 2020-06-03 10:10:54 +02:00
parent 014e70fe35
commit f893b339c5
1 changed files with 9 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

11
ppn2ocr
View file

@ -70,6 +70,14 @@ def iiif_url_for_dms_url(dms_url, ppn, size):
return iiif_url return iiif_url
def remove_file_grp(mets, use):
for bad_fileid in mets.xpath(f'//mets:fileGrp[@USE="{use}"]/mets:file/@ID', namespaces=XMLNS):
for bad in mets.xpath(f'//mets:fptr[@FILEID="{bad_fileid}"]', namespaces=XMLNS):
bad.getparent().remove(bad)
for bad in mets.xpath(f'//mets:fileGrp[@USE="{use}"]', namespaces=XMLNS):
bad.getparent().remove(bad)
def make_workspace(ppn, workspace): def make_workspace(ppn, workspace):
# Make workspace directory # Make workspace directory
os.mkdir(workspace) os.mkdir(workspace)
@ -80,8 +88,7 @@ def make_workspace(ppn, workspace):
# XXX # XXX
# Delete PRESENTATION file group # Delete PRESENTATION file group
# (local file:/// links, not handled well by "ocrd workspace") # (local file:/// links, not handled well by "ocrd workspace")
for bad in mets.xpath('//mets:fileGrp[@USE="PRESENTATION"]', namespaces=XMLNS): remove_file_grp(mets, 'PRESENTATION')
bad.getparent().remove(bad)
# Duplicate DEFAULT file group into a new file group BEST # Duplicate DEFAULT file group into a new file group BEST