Create file parsing schema

[lboeman]

We'll need a schema to communicate how to parse data from a file upload depending on what level the user provides data, e.g. system level weather or inverter level weather.

I think this will require a system-breakdown mapping and a file-header-to-variable mapping like:

Simple file-header-to-variable mapping:

{
  "file_header": "some_csv_header",
  "variable": "ghi",
}

System-breakdown describes what part of the system spec relates to what data, with a general structure that matches the System spec

{ 
  "name": "<name of system>",
  "object_type": "system",
  "mappings": [..],
  "children": [ {
     "name": "<name of inverter>", // maybe just use index.
     "object_type": "inverter",
     "mappings": [..]
     "children": [{
       "name": "<name of array>",
       "object_type": "array",
       "mappings": [...],
     },..]
   },..]
}

Example mapping for system-level data

{
  "name": "System A",
  "object_type": "system",
  "mappings": [{
      "file_header": "some_csv_header",
      "variable": "ghi",
    },
    {... },
  ]
}

Example mapping for inverter-level data

{
  "name": "System A",
  "object_type": "system",
  "mappings": [ ],
  "children": [
     {
        "name": "Inverter 1",
        "object_type": "inverter",
        "mappings: [
          {
             "file_header": "some_csv_header",
             "variable": "ghi",
          }
         ]
     },
     {...},
  ]
}

[lboeman]

It now seems like a simple mapping of schema location to mapping object will be more appropriate. Something like

{
   "/inverters/0": {
     "time": "t",
     "ghi": "global",
     "dni": "direct",
      ...
  }
}

This is what the dashboard is using for performing the mapping from user uploaded csv to the data expected by the api. Perhaps for saving a mapping we store it along with the system id and original file name so the user has an idea of what it was originally used for.