Test case tweak #349
Test case tweak #349
Conversation
So that they describe the expected behavior better.
This now correctly handles cases where there was no failure in the action. When an action fails we save the initial state that went in. However, when an action succeeds, we would only store the final state for that action. So now, if the action was successful, we go back to the prior sequence ID to get its state as the input, and use the state from the action as the desired final state. This assumes you want to say "hey here's expected behavior, I want to enshrine it". Adjusts docs with new option.
skrawcz
force-pushed
the
test_case_tweak
branch
from
114af10 to
9165661
Compare
The docs refactor missed fixing some links. This also throws in a regenerated state graph image too.
So demo data is only relevant for the local backend. It also should use the `burr_path` env variable if present. This is unclean because we leak encapsulation from the actual burr local backend...
| base_dir = os.path.expanduser("~/.burr") | ||
| if not no_copy_demo_data: | ||
|
|
||
| if backend == "local" and not no_copy_demo_data: |
There was a problem hiding this comment.
Best to instantiate the settings module if we can -- that way it's more future-proof/solves your TODO
| @@ -322,13 +333,19 @@ def import_from_file(file_path: str) -> ModuleType: | |||
| required=False, | |||
| help="Python file or fully qualified python module to import for custom serialization/deserialization.", | |||
| ) | |||
| @click.option( | |||
There was a problem hiding this comment.
Why do we need to change this? A little unclear what this does -- worth adding a bit more docs
There was a problem hiding this comment.
it provides the name of the action to test for the test case.
enables you to use output from somewhere else as input for another action easily.
| - state loaded would be the ending state, so to get the starting state, we need to look at the prior sequence ID | ||
| """ | ||
| if after_action["status"] == "completed": | ||
| print("Action was successful so loading initial and expected state into test fixture.") |
There was a problem hiding this comment.
We should have logger.warn instead of print here as we use stdout.
There was a problem hiding this comment.
prints are used throughout here.
There was a problem hiding this comment.
We shouldn't -- it sometimes will/sometimes won't parse, and you'll be putting that to a file. You want to be able to pipe it.
|
|
||
| # TODO -- make this configurable through an env variable | ||
| To override the path, set a `burr_path` environment variable to the path you want to use. |
There was a problem hiding this comment.
Point out the settings class that corresponds to this
Clarifies test case creation behavior and adds ability to override the action name
Changes
How I tested this
Notes
Checklist