`read_grammatical_expressions` will fail on grammars with lambda expressions

[Ashvin-Ranjan]

Grammars which use lambda expressions in some rules (such as the kinship grammar) will not load when read_grammatical_expressions is called upon their YAML file, with the error being that yaml.load cannot find [grammar name].<lambda>. It might be good to find a way to store these lambdas as removing them will cause load times to be increase as certain Meanings will need to be recomputed.

[Ashvin-Ranjan]

Looking at this further, it seems that using lambda expressions in general with ULTK grammars is generally bad, as they also use up a large amount of memory when generating expressions at larger depths. I have made a solution here were I instead of rules return an "ID" of the lambda expression they want and the arguments to it, and have some collapsing functions which turn the tuples into the actual outputs. It might be good to template somehow this since it is quite messy:
https://github.com/Ashvin-Ranjan/ultk/blob/kinship/src/examples/kinship/grammar/full_optimized.py