Use get_element_by_class() from utils to get rid of yet another regex.
This function used to return only the content of the element, and not
the element itself, including its tag and attributes. The whole group
of get_element_by_X() functions are a bit of a misnomer, as they all
return the *content* of the element and not the element itself.
All these functions can now return the whole element when setting
their `include_tag` parameter to `True`. By default it is `False` so
no other code will be affected by this change. Tests have been added
to test/test_utils.py accordingly.
This uncovered a bug which prevented elements starting with a hyphen as
their class name from being found. This has been fixed by fixing the
regex used in get_elements_by_class().
The condition for assigning to json_url is the wrong way round:
currently for url: aaa.com/xxx
we get:
aaa.com/xxx&output=json
instead of the correct value:
aaa.com/xxx?output=json
