akn_parser.py

from lxml import etree

ns = 'http://docs.oasis-open.org/legaldocml/ns/akn/3.0/CSD11'
ukm = 'http://www.legislation.gov.uk/namespaces/metadata'
hPath = etree.XPath('.//a:heading//text()', namespaces={'a': ns})
iPath = etree.XPath('.//a:intro//text()', namespaces={'a': ns})
cPath = etree.XPath('.//a:content//text()|.//a:content//a:ref/text()',
                    namespaces={'a': ns})
mdPath = etree.XPath('//ukm:PrimaryMetadata[1]', namespaces={'ukm': ukm})
yPath = etree.XPath('./ukm:Year/@Value', namespaces={'ukm': ukm})
nPath = etree.XPath('./ukm:Number/@Value', namespaces={'ukm': ukm})
dPath = etree.XPath('./ukm:EnactmentDate/@Date', namespaces={'ukm': ukm})
tPath = etree.XPath('./ukm:DocumentMainType/@Value', namespaces={'ukm': ukm})
sPath = etree.XPath('./ukm:DocumentStatus/@Value', namespaces={'ukm': ukm})
cPath = etree.XPath('./ukm:DocumentCategory/@Value', namespaces={'ukm': ukm})

def parse_file(fname):
    sections = {}
    context = etree.iterparse(fname, events=('start',), tag='{%s}section' % ns,
                              remove_blank_text=True, encoding='utf-8')

    def parse_sec(section):
        def make_count(xp):
            text = ' '.join(xp(section))
            return {'text': text, 'words': len(text.split()), 'chars': len(text)}
        sec = {'heading': make_count(hPath),
               'intro': make_count(iPath),
               'content': make_count(cPath)}
        for c in ['words', 'chars']:
            sec[c] = sum(sec[s][c] for s in ['heading', 'intro', 'content'])
        return sec

    for _, section in context:
        sections[section.attrib['eId']] = parse_sec(section)

    for c in ['words', 'chars']:
        sections[c] =

    count = labmda c: sum(s[c] for s in sections.values() if isinstance(s, dict))
    md = mdPath(etree.parse(fname))
    return {'sections': sections,
            'year': yPath(md),
            'number': nPath(md),
            'date': dPath(md),
            'type': tPath(md),
            'status': sPath(md),
            'category': cPath(md),
            'words': count('words'),
            'chars': count('chars')}