Avoid replacing so many characters with `_` in `_parse_raw_sam_df()`

[Peque]

As suggested by @wholmgren.

I too think it would be more convenient. For example, it would allow for easier distinction between manufacturer and module number.

[Peque]

@wholmgren @cwhanse Since we already broke the API by changing the module and inverter databases (i.e.: module/inverter names need to be updated to work with 0.7), would you be okay to introduce this change as well?

Asking before spending time preparing the PR. 😊

[Peque]

Just adding another vote in favor of this change:

@adriesse

I've never use the tab completion and find the database more convenient in its non-transposed form.

[cwhanse]

I agree with making changes to the module dataframe keys. I think our goal is to remain as consistent as possible with the source. We can propose changes to the source, if that's the best course. The SAM developers agree with the overall goal to have one name for a module that is common among both codes.

[Peque]

@cwhanse @wholmgren I would propose to have a separator between the manufacturer name and the model name, either a , (separate columns in the CSV) or some other character. It used to be the case (a : was used), but it seems not anymore. 😕

I opened this issue, in case you want to upvote/comment: NREL/SAM#227

I may wait for that to be resolved before implementing this change (feeling lazy, in case I have to change the names 2 times... 😂)

[cwhanse]

@Peque looking for closure on v0.7 release. Would you prefer to defer work in pvlib on this issue, pending any changes in SAM?

[Peque]

@cwhanse Thanks for the ping.

I would prefer to defer work pending changes in SAM, if you do not mind.

[cwhanse]

@Peque don't mind at all.

[Peque]

Maybe related (creating a separate quick-to-load Python package for parts): https://groups.google.com/forum/#!topic/pvlib-python/Sqe84M2C6RY