Fix irrelevant system error thrown in compute feature importances on LightGBM model training failure

[imatiach-msft]

Fix irrelevant system error thrown in compute feature importances on LightGBM model training failure.
Resolve customer ICM where incorrect error message was displayed due to faulty logic added in PR #2929:

The logic it should instead be equivalent to:

if "Unknown label type: 'continuous'" in str(e) or "Unknown label type: continuous" in str(e):

NOT the current format which always falls into the if as true due to string not being checked:

if "Unknown label type: 'continuous'" or "Unknown label type: continuous" in str(e):

The underlying customer error was:
ValueError: Series.dtypes must be int, float or bool
but it was re-raised as the CONTINUOUS_ERR incorrectly.

Copilot summary:
This pull request includes changes to the compute_feature_importance.py file to improve error handling for continuous target columns when creating a LightGBM model. The most important changes include the introduction of a new error variant list and the modification of the error handling logic in the create_lightgbm_model function.

Error handling improvements:

• assets/model_monitoring/components/src/feature_importance_metrics/compute_feature_importance.py: Added CONTINUOUS_SYSTEM_ERR_VARIANTS list to handle different error message formats for continuous target columns.
• assets/model_monitoring/components/src/feature_importance_metrics/compute_feature_importance.py: Updated the error handling in create_lightgbm_model to check against CONTINUOUS_SYSTEM_ERR_VARIANTS using the any function.

[github-actions]

Test Results for model-monitoring-ci

63 tests   63 ✅  54s ⏱️
 1 suites   0 💤
 1 files     0 ❌

Results for commit 232c980.

♻️ This comment has been updated with latest results.

[github-actions]

Test Results for model-monitoring-gsq-ci

13 tests   13 ✅  1h 27m 36s ⏱️
 1 suites   0 💤
 1 files     0 ❌

Results for commit 90e75c1.

♻️ This comment has been updated with latest results.