Improved hypothesis outputs

[iamdefinitelyahuman]

What I did

Improve output for hypothesis tests.

Related to #567

How I did it

1. Hypothesis has it's own version of __tracebackhide__ called __hypothesistracebackhide__ that gets applied to each brownie module to reduce the length of tracebacks.
2. Alter the hypothesis reports to include highlights and reduce file paths. This applies to "regular" hypothesis output - that of a @given based test which is shown in a === hypothesis === report.
3. Replace the default hypothesis reporter with a custom one that applies the same formatting as the previous step. This is a requirement for stateful test output, because it gets sent to stdout rather than being displayed as a report.
4. Move report_multiple_bugs to the config file so that it can be set by users. I've left it as False by default, because I feel the additional context provided by pytest's traceback rewrites is valuable enough to make it worth running a test twice. I've also added a note to the documentation about the existence of this setting and what it does.

How to verify it

Run some hypothesis tests, make them fail. This is venturing into an area where it's pretty difficult to effectively test 😬

[matnad]

lgtm