Piston syntax inserts meta-controls inside python basic objects that can be evaluated to new objects. It is intended to be used in literal markup languages such as JSON or Yaml that can then easily be parameterized.
While this approach certainly has drawbacks - mainly mixin metadata and data - it can also be very lean when JSON/Yaml is almost enough but not quite. It also fits easily in pipelines that manipulate pure JSON/Yaml, and lets you have a late evaluation instead of a templated pre-rendering.
Piston evaluates any python value.
>>> piston(0)
0
>>> piston({'some': 'python value'})
{'some': 'python value'}It can be passed a context with values to be used while expanding templates.
>>> piston({'some': 'python {value}'}, context={'value': 'fun'})
{'some': 'python fun'}Expressions are traversed through built-in collections.
>>> piston([0, {'foo': ['{i}']}, 2], context={'i': 1})
[0, {'foo': ['1']}, 2]All controls inherit from piston.Control and are documented directly
in python docstrings.
* literal strings are expanded through `str.format`.
* `$for` and `$in`: Repeat an object for each element of a collection.
* `$if`, `$then`, `$else`: Conditionally include a value or another.
* `$merge`: Merge a dictionary inside its parent.
A context can be passed containing variables accessible in python expressions.