[add data] edit pipeline

[BigFunger]

Prototype Only

• Adds converters for es_to_kibana for all processors
• Adds pipelines endpoint for listing all pipelines, and loading an existing pipeline
• Adds initial step to add data wizard that allows the user to select an existing pipeline instead of creating a new one. This step is very rough, and isn't even necessarily how I would design it. It's included to illustrate the load functionality
• Moved the logic that initializes a processor with the data from another processor. This is used when loading from storage, and when returning to the pipeline step from later steps in the wizard. This logic used to live in the pipeline addExisting function, I've moved it into the constructor of each processor. Each processor now only loads the data defined in it's model.

Note: I did not modify the later steps of the wizard that would allow the modified pipeline to be saved over an older version of itself. This functionality was initially written by @Bargs, and I would like to discuss how that would work first.

I think the changes made to the processors is worthy of it's own pull, and could go in independent of the 'load saved pipeline' functionality.

[Bargs]

Nit, but it might make intentions more clear if you use _.defaults here instead of _.assign since you're basically setting defaults.

[Bargs]

Overall the approach looks solid. I added a few inline comments. Since it's just a prototype I tried to only focus on higher level details for now (but as you can tell I couldn't help myself occasionally, sorry :) )

A few other thoughts:

• Maybe converters should contain methods for converting both ways instead of splitting them into separate modules? Could be fromKibana and toKibana maybe? I'm just thinking it might make things easier to understand.
• Regardless of how the pipeline selection ultimately gets implemented, I think the pipeline-setup directive should probably be responsible for initiating the conversion of the Kibana API pipeline into a view-model Pipeline. Outside code shouldn't have to know how to use the view-model classes specific to this directive. This basically goes back to my desire to pass pure model objects to and from these wizard step directives. Maybe with some additional thought we can make it work.
• I think we're already on the same page here, but we'll need to talk to Jurgen and maybe Court and Rashid after they've decided how the new "management" sections are going to be organized before we know exactly how this edit capability is going to be exposed.

[epixa]

@BigFunger Is this PR still relevant?