-- KDL LPeg lexer for vis.
-- Covers: // and /* */ comments, slashdash /-, quoted and raw strings,
-- type annotations (type), keywords, numbers, node/property identifiers.

local lexer = lexer
local P, S = lpeg.P, lpeg.S

local lex = lexer.new(..., {fold_by_indentation = true})

-- Comments.
local line_comment = lexer.to_eol('//')
local block_comment = lexer.range('/*', '*/', false, false, true)
lex:add_rule('comment', lex:tag(lexer.COMMENT, line_comment + block_comment))

-- Slashdash node suppression.
lex:add_rule('slashdash', lex:tag(lexer.OPERATOR, P('/-')))

-- Keywords.
lex:add_rule('keyword', lex:tag(lexer.KEYWORD, lexer.word_match({'true', 'false', 'null'})))

-- Strings: raw #"..."# first so the hash prefixes are not split off,
-- then regular "..." strings.
local raw_str = lexer.range(P('#')^1 * '"', '"' * P('#')^1)
local dq_str = lexer.range('"')
lex:add_rule('string', lex:tag(lexer.STRING, raw_str + dq_str))

-- Type annotations like (path)file or (const).
local type_ann = P('(') * lexer.word * P(')')
lex:add_rule('type', lex:tag(lexer.TYPE, type_ann))

-- Numbers (int/float/hex/oct/bin; underscores are tolerated inside).
local num_pattern =
	(lexer.number *
		(S('_')^0 * lexer.number)^0) +
	'0x' * (lpeg.R('09') + lpeg.R('af') + lpeg.R('AF') + '_')^1 +
	'0o' * lpeg.R('07')^1 +
	'0b' * lpeg.R('01')^1
lex:add_rule('number', lex:tag(lexer.NUMBER, num_pattern))

-- Identifiers (node names, property keys).
lex:add_rule('identifier', lex:tag(lexer.IDENTIFIER, lexer.word))

-- Operators / punctuation.
lex:add_rule('operator', lex:tag(lexer.OPERATOR, S('{}()=;,/+\\')))

lexer.property['scintillua.comment'] = '//'

return lex