Add `enableXMLOutput` to `TikaExtractor.extract`

[arcaputo3]

• Allow spark option as a boolean via tika.parser.enableXMLOutput
• In response to https://github.com/databrickslabs/tika-ocr/issues/45
• Works as expected on test data, but further testing may be warranted

[arcaputo3]

@aamend FYI - also seems to improve LLM use cases for, for example, tabular pdfs

[aamend]

Nice job! Thanks for the contrib