Detect automatically what is the type of document that will be parsed

[Urahara]

Today when user don't provide a type on construct we always fallback to html parser type.

parsel/parsel/selector.py

Lines 38 to 42 in 7ed4b24

	def _st(st):
	if st is None:
	return 'html'
	elif st in _ctgroup:
	return st

[Gallaecio]

I wonder if we should be using https://github.com/scrapy/xtractmime to implement this.