Sometimes, when writing tests, you want to assert that a data structure computed in a test matches some expected value, but the expected value might have components, such as times or generated ids that can't be predicted ahead of time. Options in this situation include:
- break the assertion into many assertions about various parts of the data structure.
- Use mocking/stubbing to control things like time and ids to make unpredictable parts of the data predictable.
Both of these approaches tend to be pretty tedious and potentially error prone.
The tiny testvars package provides an alternative that is much easier to use in many cases. The package provides a Vars class:
>>> import testvars
>>> vars = testvars.Vars()When Vars attributes are compared equal the first time, the comparison succeeds and alse sets their value:
>>> vars.x == 1 True
At that point, the value is set, and you can make assertions against it:
>>> vars.x 1 >>> vars.x > 0 True
Further comparisons are against this set value:
>>> vars.x == 2 False
This is useful when dealing with data structures with unpredictable parts:
>>> expected = {'id': vars.child_id, 'name':"Alex"} >>> {'id': 32, 'name':"Alex"} == expected True
It's also interesting when, while you don't know what values will be, you can make assertions about how values are related:
>>> expected = {'id': vars.parent_id, 'name':'Cas', ... 'children': [vars.child_id]} >>> expected == {'id': 42, 'name':'Cas', 'children':[32]} True >>> expected == {'id': 42, 'name':'Cas', 'children':[33]} False
>>> [vars.y, 2, 3, vars.y] == [9, 2, 3, 9] True
>>> [vars.z, 2, 3, vars.z] == [8, 2, 3, 9] False