Using namedtuple for pytest parameterized tests#

I’m writing some quite complex pytest parameterized tests this morning, and I was finding it a little bit hard to read the test cases as the number of parameters grew.

Here’s a pattern I figured out using Python’s namedtuple to make the test cases easier to read and edit:

1
from collections import namedtuple
2
import pytest
3

4
ManageTableTest = namedtuple(
5
    "ManageTableTest",
6
    (
7
        "description",
8
        "setup_post_data",
9
        "post_data",
10
        "expected_acls",
11
        "should_fail_then_succeed",
12
        "expected_audit_rows",
13
    ),
14
)
15

16
@pytest.mark.asyncio
17
@pytest.mark.parametrize(
18
    ManageTableTest._fields,
19
    (
20
        ManageTableTest(
21
            description="Group: add insert-row",
22
            setup_post_data={},
23
            post_data={"group_permissions_staff_insert-row": "on"},
24
            expected_acls=[
25
                {
26
                    "group_name": "staff",
27
                    "actor_id": None,
28
                    "action_name": "insert-row",
29
                    "database_name": "db",
30
                    "resource_name": "t",
31
                }
32
            ],
33
            should_fail_then_succeed=[
34
                dict(
35
                    actor={"id": "simon", "is_staff": True},
36
                    action="insert-row",
37
                    resource=["db", "t"],
38
                ),
39
            ],
40
            expected_audit_rows=[
41
                {
42
                    "group_name": "staff",
43
                    "actor_id": None,
44
                    "action_name": "insert-row",
45
                    "database_name": "db",
46
                    "resource_name": "t",
47
                    "operation_by": "root",
48
                    "operation": "added",
49
                }
50
            ],
51
        ),
52
    ),
53
)
54
def test_manage_table_permissions(
55
    description,
56
    setup_post_data,
57
    post_data,
58
    expected_acls,
59
    should_fail_then_succeed,
60
    expected_audit_rows,
61
):
62
    # Tests go here

You can see a more complete example here.

There are a couple of tricks here. I’m defining this namedtuple with the fields that I know I want to use in the tests:

1
ManageTableTest = namedtuple(
2
    "ManageTableTest",
3
    (
4
        "description",
5
        "setup_post_data",
6
        "post_data",
7
        "expected_acls",
8
        "should_fail_then_succeed",
9
        "expected_audit_rows",
10
    ),
11
)

I’m doing this purely to be able to use keyword arguments when defining my tests, which are much easier to read than a ordered list of arguments.

The pytest.mark.parametrize decorator is usually called with a comma-separated string of field names, which would look like this:

1
@pytest.mark.parametrize(
2
    "description,setup_post_data,post_data,expected_acls,should_fail_then_succeed,expected_audit_rows", [
3
    ManageTableTest(description="...",)
4
])
5
def test_manage_table_permissions(...)

That’s a bit ugly, and you end up duplicating the list of fields. Instead, I reuse the definition from the namedtuple like this:

1
@pytest.mark.parametrize(
2
    ManageTableTest._fields,
3
    (
4
        ManageTableTest(
5
            description="...",
6
        )
7
    )

One last trick: that description= field isn’t actually used by my code. Initially I had a code comment instead.

Then I found myself running pytest -x --pdb to drop into a debugger in a failing test but getting confused over which of the parameterized tests I was in. So I added the description argument so I can type description in the debugger and find out which test case failed.