Introducing PyMeta

I've just upload the initial release of a Python implementation of Alessandro Warth's OMeta, an object-oriented pattern matching language. Get the tarball here (or simply bzr get lp:pymeta). It's an extremely handy tool for all kinds of parsing and filtering jobs. Here's a tiny example:

>>> exampleGrammar = """
word ::= <letter>+:chars => ''.join(chars)
greeting ::= <word>:x ',' <spaces> <word>:y '!' => (x, y)
"""

Here's a breakdown of the syntax in the first rule: as you can probably tell, "word ::=" defines a parsing rule. "<letter>" calls a builtin rule that matches a single letter. The "+" works similar to its use in regexes, executing the expression before it multiple times and collecting the results into a list. ":chars" binds the result to a local variable, "chars". "=>" is used to provide a Python expression as the result of a successful parse.

The second rule strings multiple expressions together in sequence -- after matching a word, it looks for a comma and then calls the builtin rule "spaces" which matches any amount of whitespace, then another word and finally an exclamation point. Once that's all matched it returns a tuple containing the two matched words.

Using this in a Python program is simple:

>>> from pymeta.grammar import OMeta
>>> GreetingParser = OMeta.makeGrammar(exampleGrammar, {})
>>> gp = GreetingParser("Hello, world!")
>>> gp.apply("greeting")
('Hello', 'world')

In my next few posts I'll be showing more examples of what PyMeta can do and how it's different from other Python parsing tools.