Open ejsimley opened 2 years ago
Jordan is finding that a PySpark context is needed to register Python function as UDF, so we should find a way to run UDF from dbt. We can revisit this once Jordan demonstrates the UDF example.
This issue is a stretch goal for Q3 deliverable. Try to create an MVP example notebook demonstrating UDFs by end of June.
Throwing Jordan's link from Slack here for reference: https://stackoverflow.com/questions/46010682/how-to-register-custom-udf-jar-in-hivethriftserver2
We will likely want to call some OpenOA functions as UDFs from ENTR SQL. To demonstrate this, we discussed an initial implementation of the power curve fitting toolkit functions: https://github.com/entralliance/OpenOA/blob/main/operational_analysis/toolkits/power_curve/functions.py
We may also want to implement some of the filtering toolkit functions used to preprocess the SCADA data before fitting a power curve.