vampytest.assert_raises should accept a where parameter that is used as a condition.

def test_pepe():
    with vampytest.assert_raises(RuntimeError, where = lambda err: 'nue' in repr(err)):
        action()

There should be a vampytest.call_with alternative that works from a generator.

def generator():
    yield 1, 2
    yield 2, 3

@vampytest.new_stuff(generator)
def test_koishi(p0, p1):
    pass

The currently tested package should be auto detected instead of being required to manually define it.

This would allow greater flexibility when using vampytest, as you would not need to define the package directory as a parameter.

stdout and stderr should be captured. This would catch stray print and .write code. Would also catch warnings. If a test's io is captured it should show up at the end of the output above the failing ones.

vampytest.assert_instance should accept multiple types.

Sub test should show up separately on the output and also should be included in the counting like that as well.

Add a context manager to capture stdout and stderr.

def test_pepe():
    io_capture = vampytest.capture_io()
    
    with io_capture:
        pass
    
    value = io_capture.get_value()

When using vampytest.call_with and the test fails the parameters are not showing up in the output.

When dictionaries are checked for equality or inequality, and they fail then detailed difference should also be rendered.

Example:

... Normal output ...

Common items:
{
    ....,
    ....,
 }
 Different items:
 {
 +   ....
 +   ...,
 -   ...,
 -   ...,
 }

+ should mark the values that parameter_0 has and - that parameter_1 has. If a value is different in both, then first + should be rendered followed by its - counterpart.

The output should be sorted.

revert actually reverses a test's result and not reverts it. Should be renamed.