Experiment: streaming parse and eval

[daveyarwood]

Moved from alda-lang/alda#225.

An experimental idea I had when brainstorming about how to speed up the Alda parse => eval => play pipeline. See the original issue for details.

[daveyarwood]

TODO:

• Test that all the example scores sound correct with the new parser, to see if there are any bugs that the tests didn't catch.
• Merge Streaming parser #37 and release alda-core 0.2.0.
• Adjust and test with alda-repl-clj, release new version of that.
• Further testing and new release of Alda.