consider refactoring basic validations of forecast data

[elray1]

• It seems like many functions will want to do the kinds of checks that are being done here -- should we extract them into a function in hubUtils or hubValidations?
• Additionally/alternatively, should that function just accept an object that has the information about task ids, output types and type ids, and value columns as attributes (as outlined in this comment) and assume they've already been validated? Or call a function that's designed to do these validation checks based on that data structure?