dump METS files from an OAI harvest (metha-cat output after running metha-sync), with recursive METS downloads for multipart works

metha-dump.py

#!/usr/bin/env python3

import sys
from lxml import etree as ET

from ocrd_models.constants import NAMESPACES

NAMESPACES['oai'] = "http://www.openarchives.org/OAI/2.0/"

for curie in NAMESPACES:
    ET.register_namespace(curie, NAMESPACES[curie])

# can be too large for in-memory parsing:
#tree = ET.ElementTree()
#tree.parse(sys.stdin, parser=ET.ETCompatXMLParser(encoding='utf-8', recover=True))
#
#root = tree.getroot() # Records
#for record in root.iterfind('oai:record', NAMESPACES):

# so use incremental parsing instead:
def fast_iter(context, func, *args, **kwargs):
    """
    http://lxml.de/parsing.html#modifying-the-tree
    Based on Liza Daly's fast_iter
    http://www.ibm.com/developerworks/xml/library/x-hiperfparse/
    See also http://effbot.org/zone/element-iterparse.htm
    """
    for _, elem in context:
        func(elem, *args, **kwargs)
        # It's safe to call clear() here because no descendants will be
        # accessed
        elem.clear()
        # Also eliminate now-empty references from the root node to elem
        for ancestor in elem.xpath('ancestor-or-self::*'):
            while ancestor.getprevious() is not None:
                del ancestor.getparent()[0]
    del context

def write_mets(mets, name):
    #mets.write(name + '.xml', encoding='utf-8', xml_declaration=True)
    with open(name + '.xml', 'wb') as file:
        mets = ET.tostring(mets, pretty_print=True, encoding='utf-8')
        file.write(mets)

def download_mets(url):
    session = Session()
    retries = Retry(total=5, status_forcelist=[
        # only transient failures (probably too wide):
        408, 409, 412, 417, 423, 424, 425, 426, 428, 429, 440, 500, 503, 504, 509, 529, 598, 599])
    adapter = HTTPAdapter(max_retries=retries)
    session.mount('http://', adapter)
    session.mount('https://', adapter)
    response = session.get(url, timeout=3)
    response.raise_for_status()
    download = response.content
    mets = ET.fromstring(download)
    return mets

def dive_mets(mets, oai_id, level):
    if not mets:
        print("empty METS result for %s" % oai_id, file=sys.stderr)
        return
    if level > 2:
        raise Exception("unexpected depth of METS recursion for %s", oai_id)
    # assert oai_id == mets.find('./mets:dmdSec/mets:mdWrap/mets:xmlData/mods:mods/mods:recordInfo/mods:recordIdentifier[@source="http://digital.slub-dresden.de/oai/"]', namespaces=NAMESPACES).text
    logmap = mets.find('./mets:structMap[@TYPE="LOGICAL"]', namespaces=NAMESPACES)
    # as long as the deepest mets:div have mets:mptr, go recursive on them (up to 2 times)
    mptrs = logmap.xpath('.//mets:div[not(mets:div)]/mets:mptr/@xlink:href', namespaces=NAMESPACES)
    if len(mptrs):
        assert mets.find('./mets:fileSec/mets:fileGrp[@USE="DEFAULT"]', namespaces=NAMESPACES) is None, mptrs
        for url in mptrs:
            mets = download_mets(url)
            if not len(mets):
                print("empty METS result for %s" % url, file=sys.stderr)
                continue
            oai_id = mets.find('./mets:dmdSec/mets:mdWrap[@MDTYPE="MODS"]/mets:xmlData/mods:mods/mods:recordInfo/mods:recordIdentifier[@source="http://digital.slub-dresden.de/oai/"]', namespaces=NAMESPACES).text
            print("recursive  %s" % oai_id)
            dive_mets(mets, oai_id, level + 1)
    else:
        write_mets(mets, oai_id)

def process_record(record):
    oai_id = record.find('oai:header/oai:identifier', NAMESPACES).text
    print("processing %s" % oai_id)
    mets = record.find('oai:metadata/mets:mets', NAMESPACES)
    mets = ET.ElementTree(mets)
    dive_mets(mets, oai_id, 0)

if len(sys.argv) > 1:
    files = sys.argv[1:]
else:
    files = [sys.stdin.buffer]

for file_ in files:
    fast_iter(ET.iterparse(file_,
                           encoding='utf-8',
                           recover=True,
                           tag='{%s}record' % NAMESPACES['oai']),
              process_record)

now correctly handles recursive case:

• if in the logical structMap, there are terminal mets:div with mets:mptr, then instead of writing the current METS, download each of the @href METS and continue with that
• in this case, use the OAI identifier from the MODS instead of the retrieval record
• stop recursion after 2 levels, because (due to METS application profile, sections 2.1.2.2 and 2.1.3) we only expect
  • multivolume_work→multivolume_work/volume,
  • periodical→periodical/volume,
  • newspaper→newspaper/year→newspaper/month/day/issue