[Model Monitoring] Implement TDEngine read metrics API

[Eyal-Danieli]

Following #5620, this PR implements the metrics endpoints on top of the TDEngine.

In general, the implementation is quite similar to the V3IO Connector. For example, read_metrics endpoint includes:

• Type validate (metrics or results)
• Prepare SQL query with the provided metrics
• Query the DB and get a DataFrame result
• Parse the DataFrame into a list of metrics values per distinct results (using df_to_results_values & df_to_metrics_values that are now stored under the base class).

In addition, to keep consistency with the V3IO TSDB schema, we updated the application result table and added a new tag - result_kind.

Related JIRA: https://iguazio.atlassian.net/browse/ML-6557

[jond01]

Looks good, I had some minor comments.

My major question is about documentation: I think that installing TDEngine and setting the env vars has to be documented. If I want to test it myself, it would be hard to understand what I need to configure.