validation
ValidationCheckpointer
Bases: Checkpointer
Checkpointer which can be used to validate the output of a test.
__init__(savepoint_data_path, thresholds, rank)
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
savepoint_data_path
|
str
|
path to directory containing netcdf savepoint data |
required |
thresholds
|
SavepointThresholds
|
thresholds to check against |
required |
rank
|
int
|
rank of the process, needed to compare against the correct savepoint data |
required |
trial()
Context manager for a trial.
When entered, resets reference data comparison back to the start of the data.
A new context manager should entered before the code being tested is called, and exited at the end of code execution.
__call__(savepoint_name, **kwargs)
Checks the arrays passed as keyword arguments against thresholds specified.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
savepoint_name
|
str
|
name of the savepoint |
required |
**kwargs
|
ArrayLike
|
array data for variables in that savepoint |
{}
|
Raises:
| Type | Description |
|---|---|
AssertionError
|
if the thresholds on any variable are not met |