pmacs3/lex2_python.py

from lex2 import Grammar, ConstantRule, PatternRule, ContextPatternRule, RegionRule, DualRegionRule

class StringGrammar(Grammar):
    rules = [
        PatternRule(
            name=r'octal',
            pattern=r'\\[0-7]{3}',
        ),
        PatternRule(
            name=r'escaped',
            pattern=r'\\.',
        ),
        #PatternRule(
        #    name=r'format',
        #    pattern=r'%(?:\([a-zA-Z_]+\))?[-# +]*(?:[0-9]+|\*)?\.?(?:[0-9]+|\*)?[hlL]?[a-zA-Z%]',
        #),
    ]

class PythonGrammar(Grammar):
    rules = [
        PatternRule(
            name=r'functiondef',
            pattern=r'(?<=def )[a-zA-Z_][a-zA-Z0-9_]*',
        ),
        PatternRule(
            name=r'classdef',
            pattern=r'(?<=class )[a-zA-Z_][a-zA-Z0-9_]*',
        ),
        PatternRule(
            name=r'reserved',
            pattern=r'(?:True|None|False|Exception|self)(?![a-zA-Z0-9_])',
        ),
        PatternRule(
            name=r'keyword',
            pattern=r'(?:yield|while|try|return|raise|print|pass|or|not|lambda|is|in|import|if|global|from|for|finally|exec|except|else|elif|del|def|continue|class|break|assert|as|and)(?![a-zA-Z0-9_])',
        ),
        PatternRule(
            name=r"builtin",
            pattern=r'(?<!\.)(?:zip|xrange|vars|unicode|unichr|type|tuple|super|sum|str|staticmethod|sorted|slice|setattr|set|round|repr|reduce|raw_input|range|property|pow|ord|open|oct|object|max|min|map|long|locals|list|len|iter|issubclass|isinstance|int|input|id|hex|hash|hasattr|globals|getattr|frozenset|float|filter|file|execfile|eval|enumerate|divmod|dir|dict|delattr|complex|compile|coerce|cmp|classmethod|chr|callable|bool)(?![a-zA-Z0-9_])',
        ),

        PatternRule(
            name=r'methodcall',
            pattern=r'(?<=\. )[a-zA-Z_][a-zA-Z0-9_]*(?= *\()',
        ),
        PatternRule(
            name=r'functioncall',
            pattern=r'[a-zA-Z_][a-zA-Z0-9_]*(?= *\()',
        ),
        PatternRule(
            name=r'system_identifier',
            pattern=r'__[a-zA-Z0-9_]+__',
        ),
        PatternRule(
            name=r'private_identifier',
            pattern=r'__[a-zA-Z0-9_]*',
        ),
        PatternRule(
            name=r'hidden_identifier',
            pattern=r'_[a-zA-Z0-9_]*',
        ),
        PatternRule(
            name=r'identifier',
            pattern=r'[a-zA-Z_][a-zA-Z0-9_]*',
        ),
        PatternRule(
            name=r'delimiter',
            pattern=r'\(|\)|\[|\]|{|}|@|,|:|\.|`|=|;|\+=|-=|\*=|/=|//=|%=|&=|\|=|\^=|>>=|<<=|\*\*=',
        ),
        PatternRule(
            name=r"operator",
            pattern=r"\+|<>|<<|<=|<|-|>>|>=|>|\*\*|&|\*|\||/|\^|==|//|~|!=|%",
        ),

        PatternRule(
            name=r"integer",
            pattern=r"(?<![\.0-9a-zA-Z_])(?:0|[1-9][0-9]*|0[0-7]+|0[xX][0-9a-fA-F]+)[lL]?(?![\.0-9a-zA-Z_])",
        ),
        PatternRule(
            name=r"float",
            pattern=r"(?<![\.0-9a-zA-Z_])(?:[0-9]+\.[0-9]*|\.[0-9]+|(?:[0-9]|[0-9]+\.[0-9]*|\.[0-9]+)[eE][\+-]?[0-9]+)(?![\.0-9a-zA-Z_])",
        ),
        PatternRule(
            name=r"imaginary",
            pattern=r"(?<![\.0-9a-zA-Z_])(?:[0-9]+|(?:[0-9]+\.[0-9]*|\.[0-9]+|(?:[0-9]|[0-9]+\.[0-9]*|\.[0-9]+)[eE][\+-]?[0-9]+)[jJ])(?![\.0-9a-zA-Z_])",
        ),

        RegionRule(
            name=r'docstring',
            start=r'^ *(?P<tag>"""|\'\'\')',
            grammar=Grammar(),
            end=r'%(tag)s',
        ),
        RegionRule(
            name=r'tq_string',
            start=r'(?P<tag>"""|\'\'\')',
            grammar=Grammar(),
            end=r'%(tag)s',
        ),
        RegionRule(
            name=r'string',
            start=r'(?P<tag>"|\')',
            grammar=StringGrammar(),
            end=r'%(tag)s',
        ),

        PatternRule(
            name=r'comment',
            pattern=r'#.*$',
        ),
        PatternRule(
            name=r'continuation',
            pattern=r'\\$',
        ),
    ]