In Model Exploration, models with too many parameters can overflow the table name buffer

[riley-harper]

In Model Exploration, some of the names of the tables generated depend on the parameters to the models. Non-alphanumeric characters are replaced with underscores _. If the model has many complicated parameters or even a lot of whitespace in its parameter list, this can cause the table name to be longer than 128 characters, which causes Spark to throw an error.

Here's an example table name: model_eval_precision_recall_curve_random_forest__maxdepth___7___numtrees___100___threshold_ratio___1_2_. This name fits in the buffer, but adding some more parameters would likely cause an issue.