Using is_equal_to() with dicts works fine, but error messages under assertion failure are not as clear as they could be. Plus this would be super helpful when combined with #74.

Old:

assert_that({'a':1,'b':2,'c':3}).is_equal_to({'a':1,'b':2,'c':4})

Just says:

Expected <{'b': 2, 'c': 3, 'a': 1}> to be equal to <{'b': 2, 'c': 4, 'a': 1}>, but was not.

Step 1 - hide the equal bits

Should say something actually useful, like:

Expected <{.., 'c': 3}> to be equal to <{.., 'c': 4}>, but was not.

Where all the equal entries are simply hidden.

Step 2 - support deep dicts

Also, needs to support arbitrary dict depth. For example:

assert_that({'a':1,'b':{'x':2,y:3}}).is_equal_to({'a':1,'b':{'x':2,y:4}})

Should say something like:

Expected <{.., 'b': {.., 'y': 3}}> to be equal to <{.., 'b': {.., 'y': 4}}>, but was not.

Step 3 - ignore keys

Lastly, it would be cool to allow dict equality to ignore certain keys. For example:

assert_that({'a':1,'b':2,'c':3}).is_equal_to({'a':1,'b':2}, ignore='c')

Or ignore multiple keys:

assert_that({'a':1,'b':2,'c':3}).is_equal_to({'a':1}, ignore=['b','c'])

And it would be double amazing to ignore deep branches of keys. For example:

assert_that({'a':1,'b':{'x':2,y:3}}).is_equal_to({'a':1,'b':{'x':2}}, ignore='b.y')