vis
a vi-like editor based on Plan 9's structural regular expressions
git clone https://9o.is/git/vis.git
python.lua
(6186B)
1 -- Copyright 2006-2025 Mitchell. See LICENSE.
2 -- Python LPeg lexer.
3
4 local lexer = lexer
5 local token, word_match = lexer.token, lexer.word_match
6 local P, S, B = lpeg.P, lpeg.S, lpeg.B
7
8 local lex = lexer.new(..., {fold_by_indentation = true})
9
10 -- Classes.
11 lex:add_rule('classdef', lex:tag(lexer.KEYWORD, 'class') * lex:get_rule('whitespace') *
12 lex:tag(lexer.CLASS, lexer.word))
13
14 -- Keywords.
15 lex:add_rule('keyword', lex:tag(lexer.KEYWORD, lex:word_match(lexer.KEYWORD)) +
16 lex:tag(lexer.KEYWORD .. '.soft', lex:word_match(lexer.KEYWORD .. '.soft')))
17
18 -- Functions.
19 local builtin_func = -B('.') *
20 lex:tag(lexer.FUNCTION_BUILTIN, lex:word_match(lexer.FUNCTION_BUILTIN))
21 local special_func = lex:tag(lexer.FUNCTION_BUILTIN .. '.special',
22 lex:word_match(lexer.FUNCTION_BUILTIN .. '.special'))
23 local func = lex:tag(lexer.FUNCTION, lexer.word)
24 local method = B('.') * lex:tag(lexer.FUNCTION_METHOD, lexer.word)
25 lex:add_rule('function', (builtin_func + special_func + method + func) * #(lexer.space^0 * '('))
26
27 -- Constants.
28 local builtin_const = lex:tag(lexer.CONSTANT_BUILTIN, lex:word_match(lexer.CONSTANT_BUILTIN))
29 local attr = lex:tag(lexer.ATTRIBUTE, B('.') * lex:word_match(lexer.ATTRIBUTE) + '__name__')
30 lex:add_rule('constant', builtin_const + attr)
31
32 -- Strings.
33 local sq_str = lexer.range("'", true)
34 local dq_str = lexer.range('"', true)
35 local tq_str = lexer.range("'''") + lexer.range('"""')
36 lex:add_rule('string', lex:tag(lexer.STRING, (S('fFrRbBrR') * S('rRfFrRbB') + S('ruRUfFbB'))^-1 *
37 (tq_str + sq_str + dq_str)))
38
39 -- Identifiers.
40 lex:add_rule('identifier', lex:tag(lexer.IDENTIFIER, lexer.word))
41
42 -- Comments.
43 lex:add_rule('comment', lex:tag(lexer.COMMENT, lexer.to_eol('#', true)))
44
45 -- Numbers.
46 lex:add_rule('number', lex:tag(lexer.NUMBER, lexer.number_('_') * S('jJ')^-1))
47
48 -- Decorators.
49 lex:add_rule('decorator', lex:tag(lexer.ANNOTATION, '@' * lexer.word))
50
51 -- Operators.
52 lex:add_rule('operator', lex:tag(lexer.OPERATOR, S('!@%^&*()[]{}-=+/|:;.,<>~')))
53
54 -- Word lists.
55 lex:set_word_list(lexer.KEYWORD, {
56 'and', 'as', 'assert', 'async', 'await', 'break', 'class', 'continue', 'def', 'del', 'elif',
57 'else', 'except', 'False', 'finally', 'for', 'from', 'global', 'if', 'import', 'in', 'is',
58 'lambda', 'None', 'nonlocal', 'not', 'or', 'pass', 'raise', 'return', 'True', 'try', 'while',
59 'with', 'yield'
60 })
61
62 lex:set_word_list(lexer.KEYWORD .. '.soft', '_ case match')
63
64 lex:set_word_list(lexer.FUNCTION_BUILTIN, {
65 'abs', 'aiter', 'all', 'any', 'anext', 'ascii', 'bin', 'bool', 'breakpoint', 'bytearray', 'bytes',
66 'callable', 'chr', 'classmethod', 'compile', 'complex', 'delattr', 'dict', 'dir', 'divmod',
67 'enumerate', 'eval', 'exec', 'filter', 'float', 'format', 'frozenset', 'getattr', 'globals',
68 'hasattr', 'hash', 'help', 'hex', 'id', 'input', 'int', 'isinstance', 'issubclass', 'iter', 'len',
69 'list', 'locals', 'map', 'max', 'memoryview', 'min', 'next', 'object', 'oct', 'open', 'ord',
70 'pow', 'print', 'property', 'range', 'repr', 'reversed', 'round', 'set', 'setattr', 'slice',
71 'sorted', 'staticmethod', 'str', 'sum', 'super', 'tuple', 'type', 'vars', 'zip', '__import__'
72 })
73
74 lex:set_word_list(lexer.FUNCTION_BUILTIN .. '.special', {
75 '__new__', '__init__', '__del__', '__repr__', '__str__', '__bytes', '__format__', '__lt__',
76 '__le__', '__eq__', '__ne__', '__gt__', '__ge__', '__hash__', '__bool__', --
77 '__getattr__', '__getattribute__', '__setattr__', '__delattr__', '__dir__', --
78 '__get__', '__set__', '__delete__', '__slots__', --
79 '__init_subclass__', '__set_name__', --
80 '__instancecheck__', '__subclasscheck__', --
81 '__class_getitem__', --
82 '__call__', --
83 '__len__', '__length_hint', '__getitem__', '__setitem__', '__delitem__', '__missing__',
84 '__iter__', '__reversed__', '__contains__', --
85 '__add__', '__sub__', '__mul__', '__matmul__', '__truediv__', '__floordiv__', '__mod__',
86 '__divmod__', '__pow__', '__lshift__', '__rshift__', '__and__', '__xor__', '__or__', --
87 '__radd__', '__rsub__', '__rmul__', '__rmatmul__', '__rtruediv__', '__rfloordiv__', '__rmod__',
88 '__rdivmod__', '__rpow__', '__rlshift__', '__rrshift__', '__rand__', '__rxor__', '__ror__', --
89 '__iadd__', '__isub__', '__imul__', '__imatmul__', '__itruediv__', '__ifloordiv__', '__imod__',
90 '__idivmod__', '__ipow__', '__ilshift__', '__irshift__', '__iand__', '__ixor__', '__ior__', --
91 '__neg__', '__pos__', '__abs__', '__invert__', '__complex__', '__int__', '__float__', '__index__',
92 '__round__', '__trunc__', '__floor__', '__ceil__', --
93 '__enter__', '__exit__', --
94 '__match_args__', --
95 '__await__', --
96 '__aiter__', '__anext__', '__aenter__', '__aexit__' --
97 })
98
99 lex:set_word_list(lexer.CONSTANT_BUILTIN, {
100 'BaseException', 'Exception', 'Exception', 'ArithmeticError', 'BufferError', 'LookupError', --
101 'AssertionError', 'AttributeError', 'EOFError', 'FloatingPointError', 'GeneratorExit',
102 'ImportError', 'ModuleNotFoundError', 'IndexError', 'KeyError', 'KeyboardInterrupt',
103 'MemoryError', 'NameError', 'NotImplementedError', 'OSError', 'OverflowError', 'RecursionError',
104 'ReferenceError', 'RuntimeError', 'StopIteration', 'StopAsyncIteration', 'SyntaxError',
105 'IndentationError', 'TabError', 'SystemError', 'SystemExit', 'TypeError', 'UnboundLocalError',
106 'UnicodeError', 'UnicodeEncodeError', 'UnicodeDecodeError', 'UnicodeTranslateError', 'ValueError',
107 'ZeroDivisionError', --
108 'EnvironmentError', 'IOError', 'WindowsError', --
109 'BlockingIOError', 'ChildProcessError', 'ConnectionError', 'BrokenPipeError',
110 'ConnectionAbortedError', 'ConnectionRefusedError', 'FileExistsError', 'FileNotFoundError',
111 'InterruptedError', 'IsADirectoryError', 'NotADirectoryError', 'PermissionError',
112 'ProcessLookupError', 'TimeoutError', --
113 'Warning', 'UserWarning', 'DeprecationWarning', 'PendingDeprecationWarning', 'SyntaxWarning',
114 'RuntimeWarning', 'FutureWarning', 'ImportWarning', 'UnicodeWarning', 'BytesWarning',
115 'ResourceWarning'
116 })
117
118 lex:set_word_list(lexer.ATTRIBUTE, {
119 '__doc__', '__name__', '__qualname__', '__module__', '__defaults__', '__code__', '__globals__',
120 '__dict__', '__closure__', '__annotations__', '__kwdefaults__', --
121 '__file__', '__bases__', --
122 '__class__', --
123 '__self__', '__func__' --
124 })
125
126 lexer.property['scintillua.comment'] = '#'
127
128 return lex