Python-based parsing at the speed of C using Bison and Flex

Sander Mathijs van Veen 0a49121e41 Added reset_flex_buffer hook to support flushing flex's internal buffer. 13 years ago
doc 1bd7873dfa Restructured modules, fixed intra-package imports and applied pep8. 14 years ago
examples 1bd7873dfa Restructured modules, fixed intra-package imports and applied pep8. 14 years ago
src 0a49121e41 Added reset_flex_buffer hook to support flushing flex's internal buffer. 13 years ago
utils 1bd7873dfa Restructured modules, fixed intra-package imports and applied pep8. 14 years ago
.gitignore a0e0118392 removed old generated files and update git ignore list. 14 years ago
CHANGELOG a8765c6130 Initial commit of PyBison (version 0.1.8) 14 years ago
COPYING a8765c6130 Initial commit of PyBison (version 0.1.8) 14 years ago
CREDITS a8765c6130 Initial commit of PyBison (version 0.1.8) 14 years ago
INSTALL a8765c6130 Initial commit of PyBison (version 0.1.8) 14 years ago
Makefile f5301e6a1e Makefile cleanup and removed generated pxi file. 14 years ago
README a8765c6130 Initial commit of PyBison (version 0.1.8) 14 years ago
__init__.py 1bd7873dfa Restructured modules, fixed intra-package imports and applied pep8. 14 years ago
setup.py 1bd7873dfa Restructured modules, fixed intra-package imports and applied pep8. 14 years ago

README

Welcome to PyBison

Bringing GNU Bison/Flex's raw speed and power to Python

1) What is PyBison?

PyBison is a framework which effectively 'wraps' Bison and Flex into a Python
class structure.

You define a parser class, define tokens and precedences as attributes,
and parse targets as methods with rules in the docstrings,
then instantiate and run.

Black Magick happens in the background, whereupon you get callbacks each
time yyparse() resolves a parse target.

2) There are already parsers for Python. Why re-invent the wheel?

I looked at all the Python-based parsing frameworks.

IMO, the best one was PLY - a pure-python lexx/yacc implementation
(which I have borrowed from heavily in designing PyBison's OO model).

But PLY suffers some major limitations:

* usage of 'named groups' regular expressions in the lexer creates
a hard limit of 100 tokens - not enough to comfortably handle major
languages

* pure-python implementation is a convenience, but incurs a cruel
performance penalty

* the parser engine is SLR, not full LALR(1)

The other frameworks utilise a fiddly script syntax -

3) How do I use this?

Refer to the INSTALL file for setting up.
Refer to the examples and the doco for usage.