Runtime value checking for LaunchSettings and BatchSettings #740
Conversation
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## smartsim-refactor #740 +/- ##
=====================================================
- Coverage 40.45% 37.90% -2.56%
=====================================================
Files 110 109 -1
Lines 7326 6583 -743
=====================================================
- Hits 2964 2495 -469
+ Misses 4362 4088 -274
|
| :raises TypeError: if not an int | ||
| """ | ||
| if not isinstance(num_nodes, int): | ||
| raise TypeError("num_nodes argument was not of type int") | ||
|
|
||
| self.set("nodes", str(num_nodes)) |
There was a problem hiding this comment.
More philosophy than science question here: Should we check for inapropriate values here as well? Something like:
if num_nodes <= 0:
raise ValueError("Number of nodes must be a positive value")?
There was a problem hiding this comment.
A similar question exists for other places we expect integer inputs.
smartsim/settings/batch_settings.py
Outdated
| if batch_args is not None: | ||
| if not ( | ||
| isinstance(batch_args, dict) | ||
| and all(isinstance(key, str) for key, val in batch_args.items()) | ||
| ): | ||
| raise TypeError( | ||
| "batch_args argument was not of type mapping of str and str" | ||
| ) |
There was a problem hiding this comment.
Minor nit: Typehint and runtime check state that this must be a dict[str, str | None], but the error messages says it can be a Mapping[str, str]
| if batch_args is not None: | |
| if not ( | |
| isinstance(batch_args, dict) | |
| and all(isinstance(key, str) for key, val in batch_args.items()) | |
| ): | |
| raise TypeError( | |
| "batch_args argument was not of type mapping of str and str" | |
| ) | |
| if batch_args is not None: | |
| if not ( | |
| isinstance(batch_args, dict) | |
| and all(isinstance(key, str) for key, val in batch_args.items()) | |
| ): | |
| raise TypeError( | |
| "batch_args argument was not of type dict of str and str or None" | |
| ) |
There was a problem hiding this comment.
Also, since we are already iterating over batch_args, should we runtime type check the values as well as the keys of batch_args?
There was a problem hiding this comment.
the type hint for it is t.Dict[str, t.Optional[str]], so what should the value type check look like? how do I check for optional str?
smartsim/settings/batch_settings.py
Outdated
| @env_vars.setter | ||
| def env_vars(self, value: t.Dict[str, str | None]) -> None: | ||
| """Set the environment variables.""" | ||
|
|
||
| if not ( | ||
| isinstance(value, t.Mapping) | ||
| and all(isinstance(key, str) for key, val in value.items()) | ||
| ): | ||
| raise TypeError("env_vars argument was not of type dic of str and str") |
There was a problem hiding this comment.
Similar inconsistency here between dict and Mapping
There was a problem hiding this comment.
Also a similar question about whether we should runtime check the types of the values of value
smartsim/settings/launch_settings.py
Outdated
| if launch_args is not None: | ||
| if not ( | ||
| isinstance(launch_args, t.Mapping) | ||
| and all(isinstance(key, str) for key, val in launch_args.items()) | ||
| ): | ||
| raise TypeError( | ||
| "batch_args argument was not of type mapping of str and str" | ||
| ) |
There was a problem hiding this comment.
Similar 'dict vs Mapping inconsistency' and 'should we runtime type check the values of launch_args' comment here as with BatchSettings
smartsim/settings/launch_settings.py
Outdated
| @@ -165,6 +174,12 @@ def env_vars(self, value: dict[str, str | None]) -> None: | |||
|
|
|||
| :param value: The new environment mapping | |||
| """ | |||
| if not ( | |||
| isinstance(value, dict) | |||
| and all(isinstance(key, str) for key, val in value.items()) | |||
There was a problem hiding this comment.
Similar 'should we runtime type check the values of value' comment here as with BatchSettings.env_vars
| def test_type_env_vars(): | ||
| env_vars = "invalid" | ||
| with pytest.raises( | ||
| TypeError, match="env_vars argument was not of type dic of str and str" |
There was a problem hiding this comment.
| TypeError, match="env_vars argument was not of type dic of str and str" | |
| TypeError, match="env_vars argument was not of type dict of str and str" |
| ) | ||
| def test_type_launch_args(launch_args): | ||
| with pytest.raises( | ||
| TypeError, match="batch_args argument was not of type mapping of str and str" |
There was a problem hiding this comment.
| TypeError, match="batch_args argument was not of type mapping of str and str" | |
| TypeError, match="launch_args argument was not of type mapping of str and str" |
| def test_type_env_vars(): | ||
| env_vars = "invalid" | ||
| with pytest.raises( | ||
| TypeError, match="env_vars argument was not of type dic of str and str" |
There was a problem hiding this comment.
| TypeError, match="env_vars argument was not of type dic of str and str" | |
| TypeError, match="env_vars argument was not of type dict of str and str" |
No description provided.