Don't use load ID's to set priority

[bluejuniper]

trafficstars

A commonly used dataset has load id's starting with "h" and "l" Suggest keeping load priorities in a separate file or adding a new column to the loads table if the parser supports it

[ccoffrin]

In the interest of simplicity, what about changing the convention to some special characters? Say !, *, and -?

[bluejuniper]

The preference between Mike & I is to avoid reading load ids unless we know for sure that (1) special characters won't appear in them and (2) that both positions in the load id don't need to be unique. The PTI RAW33 specification is vague about (1), just saying that it's an alphanumeric field. For our use keeping a separate file of load priorities isn't a big deal since we already need to keep multiple files to record substation connectivity. I realize that adding options is a slippery slope, but I'd suggest having the parser not read load priorities by default and to do so only if a keyword option is set.

[ccoffrin]

Ok based on your suggestion I propose to remove the feature for setting load priorities in the .raw file. The load weight parameter will still be supported internally. It will be the job of the user to find some way to setting it appropriately for a specific application and source data. Does that sound suitable?

[bluejuniper]

That sounds reasonable, thanks Carleton

[ccoffrin]

I just noticed that this feature is currently exposed as a post-processing function to the user. In v0.3 simply don't call the function add_load_weights! and you should get the behavior you suggest. Nonetheless we will drop add_load_weights! in v0.4.