Define YODA format in terms of YAML base syntax

[buckley]

Since YAML is nice & clean, since we're using it elsewhere, and since making a parser for the "dat" format was proving to be a pain, we'll use YAML as the base format for YODA objects. The translation shouldn't be too hard, but this ticket can be used for kicking about some ideas. I've mocked up some examples on my laptop, which I'll copy on here shortly.

The main issues I see as up for discussion are how to represent more complex errors (which was anyway an unconsidered and awkward issue for the dat format), and whether to treat each object as a separate YAML document with the type as the top-level key, or to put them all in one doc with the path as the top key. I'm inclined to the latter, but I'll let you decide once you've seen the mock-ups...

[buckley]

A schematic example for comments: issues are

• use separate docs for each object?
• top-level key is type or path?
• how to combine weights and extra errors?

[NB. the values are represented as "variables" like xe+ rather than numbers to hopefully illustrate what they mean. Mentally replace them with doubles]

/plot/path:
  Type: PLOT
  DrawOnly: [/REF, /foo, /bar, /baz]
  RatioPlot: true
  LogY: true
  ...
---
myfile/histo/path:
  Type: HISTO
  Data:
   - [xl, xh, y, {stat:[ye-, ye+], sys:ye%}, w, w2, wx, wx2]
   - [xl, xh, y, [ye-, ye+], w, w2, wx, wx2]
   - [xl, xh, y, ye, w, w2, wx, wx2]
   - ...
---
myfile/histo/path:
  Type: PROFILE
  Data:
   - [xl, xh, y, {stat:[ye-, ye+], sys:ye}, w, w2, wx, wx2, wy, wy2]
   - [xl, xh, y, [ye-, ye+], w, w2, wx, wx2, wy, wy2]
   - [xl, xh, y, ye, w, w2, wx, wx2, wy, wy2]
   - ...
  SysErrors: [[yesys-, yesys+], [yesys-, yesys+]]
---
myfile/histo/path:
  Type: SCATTER
  Dims: 2
  Data: 
   - [x, xe-, xe+, y, ye-, ye+]
   - [x, xe-, xe+, y, ye-, ye+]
   - ...
  MoreErrors: ?
---
HISTO
FUNCTION
SPECIAL
CONFIG

[buckley]

We didn't go this way in the end...