[Pyparsing] Bringing together ANTLR and PyParsing

[Athanasios A. <ath...@gm...>]

Hello everybody

I have been using PyParsing for some time now and after i came across
ANTLR, i thought about bringing the two together. That is, specify a parser
in ANTLR's meta-language and then have that translated to a set of
PyParsing objects.

If you have not seen ANTLR before, it looks like this:

input: list+;
list: SYM_SBL atom (SYM_COMA atom)* SYM_SBR; //Tokens and rules separated
by comas effectively means concatenation.
atom:number|string; //The usual OR
number:NUM; //number is a rule, NUM is a token (Anything that starts with a
capital letter is a token else is a rule)
string:STR;

SYM_SBL:'[';
SYM_RBL:']';
SYM_COMA:',';
NUM:[0-9]+;
STR:'\'' .*? '\'';

A preliminary attempt at this transformation can be found at:
https://bitbucket.org/aanastasiou/antlr2pyparsing

The relevant discussion over at ANTLR's group is available at:
https://groups.google.com/forum/#!topic/antlr-discussion/feRPZhfMcpU

ANTLR will be generating parsers in Python soon as it seems. The relevant
discussion is available at:
https://groups.google.com/forum/#!topic/antlr-discussion/sFH5Y0QO4HA But in
any case, this transformation would be very useful to have for PyParsing
too.

I am just wondering if we could perhaps even augment ANTLR's syntax to
accommodate PyParsing features (like Group or Suppress) that ANTLR does not
know about.

What do you think?

All the best
AA