3RNN/Lib/site-packages/pygments/lexers/lean.py
2024-05-26 19:49:15 +02:00

242 lines
8.4 KiB
Python
Raw Permalink Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

"""
pygments.lexers.lean
~~~~~~~~~~~~~~~~~~~~
Lexers for the Lean theorem prover.
:copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
:license: BSD, see LICENSE for details.
"""
import re
from pygments.lexer import RegexLexer, words, include
from pygments.token import Comment, Operator, Keyword, Name, String, \
Number, Generic, Whitespace
__all__ = ['Lean3Lexer', 'Lean4Lexer']
class Lean3Lexer(RegexLexer):
"""
For the Lean 3 theorem prover.
"""
name = 'Lean'
url = 'https://leanprover-community.github.io/lean3'
aliases = ['lean', 'lean3']
filenames = ['*.lean']
mimetypes = ['text/x-lean', 'text/x-lean3']
version_added = '2.0'
# from https://github.com/leanprover/vscode-lean/blob/1589ca3a65e394b3789409707febbd2d166c9344/syntaxes/lean.json#L186C20-L186C217
_name_segment = (
"(?![λΠΣ])[_a-zA-Zα-ωΑ-Ωϊ-ϻἀ-῾℀-⅏𝒜-𝖟]"
"(?:(?![λΠΣ])[_a-zA-Zα-ωΑ-Ωϊ-ϻἀ-῾℀-⅏𝒜-𝖟0-9'ⁿ-₉ₐ-ₜᵢ-ᵪ])*")
_name = _name_segment + r"(\." + _name_segment + r")*"
tokens = {
'expression': [
(r'\s+', Whitespace),
(r'/--', String.Doc, 'docstring'),
(r'/-', Comment, 'comment'),
(r'--.*?$', Comment.Single),
(words((
'forall', 'fun', 'Pi', 'from', 'have', 'show', 'assume', 'suffices',
'let', 'if', 'else', 'then', 'in', 'with', 'calc', 'match',
'do'
), prefix=r'\b', suffix=r'\b'), Keyword),
(words(('sorry', 'admit'), prefix=r'\b', suffix=r'\b'), Generic.Error),
(words(('Sort', 'Prop', 'Type'), prefix=r'\b', suffix=r'\b'), Keyword.Type),
(words((
'(', ')', ':', '{', '}', '[', ']', '', '', '', '', '', '', ':=', ',',
)), Operator),
(_name, Name),
(r'``?' + _name, String.Symbol),
(r'0x[A-Za-z0-9]+', Number.Integer),
(r'0b[01]+', Number.Integer),
(r'\d+', Number.Integer),
(r'"', String.Double, 'string'),
(r"'(?:(\\[\\\"'nt])|(\\x[0-9a-fA-F]{2})|(\\u[0-9a-fA-F]{4})|.)'", String.Char),
(r'[~?][a-z][\w\']*:', Name.Variable),
(r'\S', Name.Builtin.Pseudo),
],
'root': [
(words((
'import', 'renaming', 'hiding',
'namespace',
'local',
'private', 'protected', 'section',
'include', 'omit', 'section',
'protected', 'export',
'open',
'attribute',
), prefix=r'\b', suffix=r'\b'), Keyword.Namespace),
(words((
'lemma', 'theorem', 'def', 'definition', 'example',
'axiom', 'axioms', 'constant', 'constants',
'universe', 'universes',
'inductive', 'coinductive', 'structure', 'extends',
'class', 'instance',
'abbreviation',
'noncomputable theory',
'noncomputable', 'mutual', 'meta',
'attribute',
'parameter', 'parameters',
'variable', 'variables',
'reserve', 'precedence',
'postfix', 'prefix', 'notation', 'infix', 'infixl', 'infixr',
'begin', 'by', 'end',
'set_option',
'run_cmd',
), prefix=r'\b', suffix=r'\b'), Keyword.Declaration),
(r'@\[', Keyword.Declaration, 'attribute'),
(words((
'#eval', '#check', '#reduce', '#exit',
'#print', '#help',
), suffix=r'\b'), Keyword),
include('expression')
],
'attribute': [
(r'\]', Keyword.Declaration, '#pop'),
include('expression'),
],
'comment': [
(r'[^/-]+', Comment.Multiline),
(r'/-', Comment.Multiline, '#push'),
(r'-/', Comment.Multiline, '#pop'),
(r'[/-]', Comment.Multiline)
],
'docstring': [
(r'[^/-]+', String.Doc),
(r'-/', String.Doc, '#pop'),
(r'[/-]', String.Doc)
],
'string': [
(r'[^\\"]+', String.Double),
(r"(?:(\\[\\\"'nt])|(\\x[0-9a-fA-F]{2})|(\\u[0-9a-fA-F]{4}))", String.Escape),
('"', String.Double, '#pop'),
],
}
def analyse_text(text):
if re.search(r'^import [a-z]', text, re.MULTILINE):
return 0.1
LeanLexer = Lean3Lexer
class Lean4Lexer(RegexLexer):
"""
For the Lean 4 theorem prover.
"""
name = 'Lean4'
url = 'https://github.com/leanprover/lean4'
aliases = ['lean4']
filenames = ['*.lean']
mimetypes = ['text/x-lean4']
version_added = '2.18'
# same as Lean3Lexer, with `!` and `?` allowed
_name_segment = (
"(?![λΠΣ])[_a-zA-Zα-ωΑ-Ωϊ-ϻἀ-῾℀-⅏𝒜-𝖟]"
"(?:(?![λΠΣ])[_a-zA-Zα-ωΑ-Ωϊ-ϻἀ-῾℀-⅏𝒜-𝖟0-9'ⁿ-₉ₐ-ₜᵢ-ᵪ!?])*")
_name = _name_segment + r"(\." + _name_segment + r")*"
keywords1 = (
'import', 'unif_hint',
'renaming', 'inline', 'hiding', 'lemma', 'variable',
'theorem', 'axiom', 'inductive', 'structure', 'universe', 'alias',
'#help', 'precedence', 'postfix', 'prefix',
'infix', 'infixl', 'infixr', 'notation', '#eval',
'#check', '#reduce', '#exit', 'end', 'private', 'using', 'namespace',
'instance', 'section', 'protected',
'export', 'set_option', 'extends', 'open', 'example',
'#print', 'opaque',
'def', 'macro', 'elab', 'syntax', 'macro_rules', '#reduce', 'where',
'abbrev', 'noncomputable', 'class', 'attribute', '#synth', 'mutual',
'scoped', 'local',
)
keywords2 = (
'forall', 'fun', 'obtain', 'from', 'have', 'show', 'assume',
'let', 'if', 'else', 'then', 'by', 'in', 'with',
'calc', 'match', 'nomatch', 'do', 'at',
)
keywords3 = (
# Sorts
'Type', 'Prop', 'Sort',
)
operators = (
'!=', '#', '&', '&&', '*', '+', '-', '/', '@', '!',
'-.', '->', '.', '..', '...', '::', ':>', ';', ';;', '<',
'<-', '=', '==', '>', '_', '|', '||', '~', '=>', '<=', '>=',
'/\\', '\\/', '', 'Π', 'λ', '', '', '', '', '', '',
'¬', '⁻¹', '', '', '', '', '', '×', '',
'', '', '', '', "",
)
punctuation = ('(', ')', ':', '{', '}', '[', ']', '', '',
':=', ',')
tokens = {
'expression': [
(r'\s+', Whitespace),
(r'/--', String.Doc, 'docstring'),
(r'/-', Comment, 'comment'),
(r'--.*$', Comment.Single),
(words(keywords3, prefix=r'\b', suffix=r'\b'), Keyword.Type),
(words(('sorry', 'admit'), prefix=r'\b', suffix=r'\b'), Generic.Error),
(words(operators), Name.Builtin.Pseudo),
(words(punctuation), Operator),
(_name_segment, Name),
(r'``?' + _name, String.Symbol),
(r'(?<=\.)\d+', Number),
(r'(\d+\.\d*)([eE][+-]?[0-9]+)?', Number.Float),
(r'\d+', Number.Integer),
(r'"', String.Double, 'string'),
(r'[~?][a-z][\w\']*:', Name.Variable),
(r'\S', Name.Builtin.Pseudo),
],
'root': [
(words(keywords1, prefix=r'\b', suffix=r'\b'), Keyword.Namespace),
(words(keywords2, prefix=r'\b', suffix=r'\b'), Keyword),
(r'@\[', Keyword.Declaration, 'attribute'),
include('expression')
],
'attribute': [
(r'\]', Keyword.Declaration, '#pop'),
include('expression'),
],
'comment': [
# Multiline Comments
(r'[^/-]+', Comment.Multiline),
(r'/-', Comment.Multiline, '#push'),
(r'-/', Comment.Multiline, '#pop'),
(r'[/-]', Comment.Multiline)
],
'docstring': [
(r'[^/-]+', String.Doc),
(r'-/', String.Doc, '#pop'),
(r'[/-]', String.Doc)
],
'string': [
(r'[^\\"]+', String.Double),
(r'\\[n"\\\n]', String.Escape),
('"', String.Double, '#pop'),
],
}
def analyse_text(text):
if re.search(r'^import [A-Z]', text, re.MULTILINE):
return 0.1