Basic rule evaluation

[pieterbos]

Basic rule evaluation. Evaluates:

• arithmetic operators
• boolean operators
• constants
• variable declarations, expression typed (not yet builtin or queries)
• variable references
• model references with absolute paths
• stores assertion results so error messages can be displayed

Interprets everything as a list, so if you do

simple_check: /data[id5]/items[id2]/value[id4]/magnitude * 3 > 5

And the ModelReference/adl_path returns two values (very very possible in the RM!), it will do both the computation and the assertion twice. If one of the two checks fail, the assertion will have failed.

The individual check results remain, so if one succeeds and one fails, the assertion result will be:

"result": false,
"rawResult": [false, true],
"ag": "simple_check"

It would be great to store all object references used in every computation, so in a UI you could show which exact fields make this assertion fail, and which are fine.

Assertion labels/tags are still odd: you can add them, and they do help, but there's no way to add to the terminology yet. I think there should be, so you can add error messages that make sense.