You cannot select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
ocrd-galley/ppn2ocr

148 lines
4.2 KiB
Plaintext

#!/usr/bin/env python3
"""Get OCR results as a OCR-D workspace for a given PPN"""
import os
import requests
import sys
import lxml.etree as ET
import re
import subprocess
import click
from copy import deepcopy
XMLNS = {
'mets': 'http://www.loc.gov/METS/',
'xlink': 'http://www.w3.org/1999/xlink'
}
API_URL = 'https://digital.staatsbibliothek-berlin.de/oai'
IDENTIFIER_TEMPLATE = 'oai:digital.staatsbibliothek-berlin.de:%s'
for prefix, uri in XMLNS.items():
ET.register_namespace(prefix, uri)
def oai_mets(ppn):
"""Retrieve METS metadata for a given PPN."""
params = {
'verb': 'GetRecord',
'metadataPrefix': 'mets',
'identifier': IDENTIFIER_TEMPLATE % ppn
}
s = requests.Session()
# FIXME oai.sbb.berlin fails certificate check
#r = s.get(API_URL, params=params)
r = s.get(API_URL, params=params, verify=False)
mets = ET.XML(r.content).find(f".//{{{XMLNS['mets']}}}mets")
mets = ET.ElementTree(mets)
return mets
def iiif_url_for_dms_url(dms_url, ppn, size):
"""
Construct an IIIF URL from a dms URL.
This function exists to encapsulate the hack of rewriting the URL to get IIIF.
"""
if ppn not in dms_url:
raise ValueError(f"Unexpected URL {dms_url}")
m = re.search(r'/dms/.*/([0-9]+)\.jpg$', dms_url)
if m:
page_num = m.group(1)
else:
raise ValueError(f"Unexpected URL {dms_url}")
iiif_identifier = f'{ppn}-{page_num}'
iiif_url = f'https://content.staatsbibliothek-berlin.de/dc/{iiif_identifier}/full/{size}/0/default.jpg'
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):
# Make workspace directory
os.mkdir(workspace)
os.chdir(workspace)
mets = oai_mets(ppn)
# XXX
# Delete PRESENTATION file group
# (local file:/// links, not handled well by "ocrd workspace")
remove_file_grp(mets, 'PRESENTATION')
# Duplicate DEFAULT file group into a new file group BEST
file_grp_default = mets.find('//mets:fileGrp[@USE="DEFAULT"]', namespaces=XMLNS)
file_grp_best = deepcopy(file_grp_default)
file_grp_best.attrib['USE'] = 'BEST'
for f in file_grp_best.findall('./mets:file', namespaces=XMLNS):
old_id = f.attrib['ID']
new_id = re.sub('DEFAULT', 'BEST', old_id)
f.attrib['ID'] = new_id
for fptr in mets.findall(f'//mets:fptr[@FILEID="{old_id}"]', namespaces=XMLNS):
new_fptr = deepcopy(fptr)
new_fptr.attrib['FILEID'] = new_id
fptr.getparent().append(new_fptr)
# XXX Need to fumble around with the URL for now
flocat = f.find(f".//{{{XMLNS['mets']}}}FLocat")
old_url = flocat.attrib[f"{{{XMLNS['xlink']}}}href"]
url_iiif_full = iiif_url_for_dms_url(old_url, ppn, 'full')
flocat.attrib[f"{{{XMLNS['xlink']}}}href"] = url_iiif_full
mets.find('//mets:fileSec', namespaces=XMLNS).append(file_grp_best)
# Write mets.xml
mets.write('mets.xml', pretty_print=True)
# TODO
# Validate workspace
#ocrd workspace validate mets.xml | grep -v "<notice>Won't download remote image"
# XXX
# Fix 'file:/' URLs to 'file:///'
#sed -i 's#file:/\([^/]\)#file:///\1#' mets.xml
# Patch mets.xml to use our NFS mount
#sed -i 's#file:///goobi/tiff001/sbb/#file:///srv/digisam_images/sbb/#g' mets.xml
@click.command()
@click.argument('ppn')
def ppn2ocr(ppn):
"""
Get METS with best images for a document PPN
For example, to get the document "PROPOSITIONES PHILOSOPHICAE: [...]" use this:
\b
ppn2ocr PPN699887615
ls PPN699887615
"""
self_dir = os.path.realpath(os.path.dirname(sys.argv[0]))
make_workspace(ppn, ppn)
# XXX
# subprocess.run([
# os.path.join(self_dir, 'run-docker-hub'),
# '-I', 'BEST',
# '--skip-validation'
# ])
if __name__ == '__main__':
ppn2ocr()