Align run name strategy between `MLflowCallback` and `track_in_mlflow` method

[TTRh]

Motivation

The strategy for infering the run name was not aligned between the base class MLflowCallback and the decorator track_in_mlflow.

• in MLflowCallback we take the run name given by the user and we fallback to trial number if not given : https://github.com/optuna/optuna-integration/blob/main/optuna_integration/mlflow.py#L149
• in decorator we take by default trial.number: https://github.com/optuna/optuna-integration/blob/main/optuna_integration/mlflow.py#L211

Description of the changes

This PR aligned the strategy to get the run name from MLfowCallback and pass it to mlflow. We lookup _mlflow_kwargs and we fallback on trial.number if not given.

[TTRh]

Tests are failing but i'm not sure it's related to my change i see in the logs:

ERROR collecting tests/lightgbm_tuner_tests/test_alias.py

[HideakiImamura]

Sorry for the confusion. I will fix the main branch. Please wait a minuite.

[HideakiImamura]

The main branch has been fixed in https://github.com/optuna/optuna-integration/pull/112. Could you merge the main branch?

[github-actions[bot]]

This pull request has not seen any recent activity.