feature/samples_summary_failsafe

[Jammy2211]

This pull request updates the behavior of the instance method in autofit/aggregator/search_output.py to improve handling of cases where the max_log_likelihood method is unavailable. Instead of returning None, it now falls back to returning the instance from samples_summary.

Error handling improvement:

• Updated the instance method to return self.samples_summary.instance when max_log_likelihood is not implemented or accessible, rather than returning None.

[Copilot]

Pull request overview

This pull request updates the instance property in the SearchOutput class to improve error handling when the max_log_likelihood method is unavailable. Instead of returning None, it now attempts to fall back to samples_summary.instance as an alternative source for the instance data.

Changes:

• Modified exception handling in the instance property to return self.samples_summary.instance instead of None when AttributeError or NotImplementedError is raised

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

This change introduces a potential AttributeError that is not caught. The samples_summary property (line 227) calls self.value("samples_summary") which can return None if the file doesn't exist. When summary is None, line 228 will fail with AttributeError: 'NoneType' object has no attribute 'model'. Additionally, even if samples_summary is not None, accessing its instance property could raise exceptions since it internally calls max_log_likelihood().

The docstring on line 248 states "None if samples cannot be loaded", but the new code path doesn't handle these potential failures. Consider wrapping the fallback in a try-except block similar to the pattern used in autofit/non_linear/result.py:115-120, or add proper null checking.