The features metrics pipeline running with the cron tasks isn't implemented as intended

[sinnec]

The features metrics pipeline running with the cron tasks isn't implemented as intended. It uses inferences while it should be using training datasets and should run when a training dataset is inserted and not in cron tasks.

Note: The current implementation could be used for data quality metrics.

Tasks:

• [ ] Calculate feature metrics when a training dataset is inserted.
• [ ] Rename all model integrity functions, properties etc so the use the same name (model integrity metrics, feature metrics or descriptive statistics)

[stavrostheocharis]

For now, let’s use it for processed training data. For inferences let’s keep it for a new feature.

[stavrostheocharis]

For the name let's use the "descriptive statistics" where we use integrity, etc.: