Can't pass DTD resolvers to LxmlEventHandler

[RobertBaruch]

Possibly related to #769.

Generated the models using generate voko/samples --config voko/samples/.xsdata.xml --output dataclasses.

Then ran abak.xml through basic parsing:

import pprint

from voko.samples.models import Vortaro
from xsdata.formats.dataclass.parsers import XmlParser
from xsdata.formats.dataclass.parsers.config import ParserConfig
from xsdata.formats.dataclass.parsers.handlers import LxmlEventHandler

if __name__ == '__main__':
    file_path = "voko/samples/abak.xml"
    config = ParserConfig(load_dtd=True)
    parser = XmlParser(config=config, handler=LxmlEventHandler)

    with open(file_path, "rb") as f:
        data = parser.parse(f, Vortaro)

    pprint.pprint(parser.parse(file_path))

In abak.xml, there is this fragment:

  <snc mrk="abak.0o.kalkulilo">
    <uzo tip="fak">MAT</uzo>
    <fnt><bib>PV</bib></fnt>
    <dif>
      <ref tip="super" cel="kalkul.0ilo">Kalkulilo</ref>
      por plenumi aritmetikajn operaciojn,
      konsistanta &ccirc;efe el buletoj &scirc;oveblaj sur
      stangetoj:
      <ekz>
      ...
      </ekz>
      ...
    </dif>
    ...
  </snc>

The Dif here is parsed as:

Dif(lng=None,
    content=[Ref(tip=<RefTip.SUPER: 'super'>,
                 val=None,
                 lst=None,
                 cel='kalkul.0ilo',
                 content=['Kalkulilo',
                          '\n'
                          '      por plenumi aritmetikajn operaciojn,\n'
                          '      konsistanta ']),
             Ekz(mrk=None,
                 content=['\n\tvi mispu',
                          Tld(lit=None, var=None),
                          'o\n\t',
                          Fnt(content=[None,
                                       Bib(value='RugxDom'),
                                       Lok(content=[]),
                                       ';\n      '])]),
             Ekz(mrk=None,
                 content=['\n        en sia animo li restis ',
                          Tld(lit=None, var=None),
                          'o\n        ',
                          Fnt(content=[None,
                                       Vrk(content=[Url(ref='https://esperanto.mv.ru/TRIZ/pretere.html',
                                                        value='Ni iros '
                                                              'pretere ')]),
                                       '.\n      '])])])

There are two errors here, possibly related.

The first error is that while Kalkulilo is the content of the <ref>, the rest should be in the content of the <dif>, not the <ref>.

The second error is that everything starting from ĉ is ignored.

It should look something like this:

Dif(lng=None,
    content=[Ref(tip=<RefTip.SUPER: 'super'>,
                 val=None,
                 lst=None,
                 cel='kalkul.0ilo',
                 content=['Kalkulilo']),
              '\n',
              '      por plenumi aritmetikajn operaciojn,\n',
              '      konsistanta ĉefe el buletoj ŝoveblaj sur\n',
              '      stangetoj:\n',
              '      ',
              Ekz(mrk=None, .....

I think the problem here is that the lxml parser isn't using the DTD because it can't find it from the xml file:

<!DOCTYPE vortaro SYSTEM "../dtd/vokoxml.dtd">

Reading through the lxml docs, I found that you have to tell it how to resolve the DTD's URI:

>>> parser = etree.XMLParser(load_dtd=True)
>>> parser.resolvers.add( DTDResolver() )

But I don't see a way to pass a resolver instance. I'd expect it to be in ParserConfig.