Adds better support for Typst

lexers/typst.lua

@@ -0,0 +1,125 @@
+local lexer = require('lexer')
+local token = lexer.token
+local P, S, B = lpeg.P, lpeg.S, lpeg.B
+
+local lex = lexer.new(...)
+-- Keep things simple for now and only allow bold and italic in non-code mode
+local italic = -B('\\') * lex:tag(lexer.ITALIC, lexer.range('_', '_'))
+local bold = -B('\\') * lex:tag(lexer.BOLD, lexer.range('*', '*'))
+
+lex:add_rule('bold', bold)
+lex:add_rule('italic', italic)
+
+local function header(level)
+  local hspace = (lexer.space - '\n')
+  local equals_signs = P('=')^level
+  -- Stupid header rule for now
+  local header = (lexer.starts_line(hspace^0 * equals_signs * hspace^1) * (lexer.any - S('\n'))^0)
+--[[
+  local header = (lexer.starts_line(hspace^0 * equals_signs * hspace^1) * (lexer.any - S('#@<\n'))^0) +
+				(((B('[') * hspace^0 * equals_signs * hspace^1)) *
+				(lexer.any - S('#@<'))^0)
+]]
+  return lex:tag(string.format('%s.h%s', lexer.HEADING, level), header)
+end
+
+local function build_rules(pre)
+  local hash_word = -B('\\') * pre * lexer.word
+  local keyword_match = -B('\\') * pre * lex:word_match(lexer.KEYWORD)
+
+  return {
+    in_code = -B('\\') * lexer.range('`', false, false),
+    dq_string = -B('\\') * lexer.range('"', true),
+    string = -B('\\') * lexer.range('`', false, false) + -B('\\') * lexer.range('"', true),
+
+    hash_word = hash_word,
+    keyword_match = keyword_match,
+
+--    TODO: limit numeric values to only be tagged when used as args, assigned values
+--    numeric_value = (lexer.number^1 * ('.' * lexer.number^1)^-1 * lex:word_match('UNITS')^-1),
+    iden = lex:tag(lexer.IDENTIFIER, hash_word),
+    mod_func = lex:tag(lexer.KEYWORD, hash_word) * lexer.space^1 * 
+               lex:tag(lexer.FUNCTION, lexer.word) * lex:tag(lexer.OPERATOR, S('[(')),
+    func = lex:tag(lexer.FUNCTION, hash_word) * lex:tag(lexer.OPERATOR, S('[(')),
+    method = lex:tag(lexer.IDENTIFIER, hash_word) *
+             lex:tag(lexer.OPERATOR, P('.')) *
+             lex:tag(lexer.FUNCTION_METHOD, lexer.word) * lex:tag(lexer.OPERATOR, S('[(')),
+    field = lex:tag(lexer.IDENTIFIER, hash_word) *
+            lex:tag(lexer.OPERATOR, P('.')) *
+            lex:tag('FIELD', lexer.word) * -S('[('),
+    operator = lex:tag(lexer.OPERATOR, S('+-/*%<>~!=^&|?~:;,.()[]{}')),
+    label = -B('\\') * lex:tag(lexer.LABEL, P('<') * lexer.word * P('>')),
+    label_two = -B('\\') * lex:tag(lexer.LABEL, P('@') * lexer.word),
+    link = P('http') * P('s')^-1 * P(':') * (lexer.word + S('.:/'))^1,
+
+    math = -B('\\') * lexer.range('$', false, false),
+    code = lexer.range('```', '```', false),
+    list = lex:tag(lexer.LIST, lexer.starts_line(lexer.digit^1 * '.' + S('+-'), true) * S(' \t')),
+    -- TODO: Do we really need to not tag a number if procceded by an alpha
+    comment = lex:tag(lexer.COMMENT, lexer.range('/*', '*/') + lexer.to_eol('//')),
+
+    keyword = lex:tag(lexer.KEYWORD, keyword_match),
+
+    header = header(6) + header(5) + header(4) + header(3) + header(2) + header(1)
+  }
+end
+
+local emb_lex = lexer.new('scripting')
+
+--[[
+ #{ ... }
+   OR
+ #let x = { ... }
+]]
+-- This is very limited, since it would only work correctly if no nested structures (w/ brackets) are found inside
+-- otherwise (if they're found inside), the first closing bracket of that nested structure would close the whole embedded
+-- script, causing the rest of the script to not be treated as a part of embedded script
+local start = (lex:tag(lexer.KEYWORD, P('#') * lex:word_match(lexer.KEYWORD)) *
+	      #((lexer.any - S('{;\n'))^1 * S('{') * lexer.space^0)) +
+	      lex:tag(lexer.OPERATOR,P('#') * S('{'))
+local embed_start = lex:tag('emb_tag', start)
+local embed_end = lexer:tag('emb_tag', S('}'))
+
+local function add_rules(lexer_obj, pre)
+  local rules = build_rules(pre)
+  lexer_obj:add_rule('header', rules.header)
+  lexer_obj:add_rule('field', rules.field)
+  lexer_obj:add_rule('function', rules.mod_func + rules.func)
+  lexer_obj:add_rule('method', rules.method)
+  lexer_obj:add_rule('label', rules.label + rules.label_two)
+  lexer_obj:add_rule('code', lex:tag(lexer.CODE, rules.code))
+  lexer_obj:add_rule('string', lex:tag(lexer.STRING, rules.string))
+  lexer_obj:add_rule('link', lex:tag(lexer.LINK, rules.link))
+  lexer_obj:add_rule('math', lex:tag('environment.math', rules.math))
+  lexer_obj:add_rule('keyword', rules.keyword)
+  lexer_obj:add_rule('identifier', rules.iden)
+  --lexer_obj:add_rule('number', lex:tag(lexer.NUMBER, rules.numeric_value))
+  lexer_obj:add_rule('list', rules.list)
+  lexer_obj:add_rule('comment', rules.comment)
+  lexer_obj:add_rule('operator', rules.operator)
+end
+
+-- Keywords, functions... don't need '#' when in code
+-- the character `#` is not valid in code
+-- TODO: only enable styling and text related rules when in []
+add_rules(emb_lex, '')
+
+lex:embed(emb_lex, embed_start, embed_end)
+
+add_rules(lex, '#')
+
+
+lex:set_word_list(lexer.KEYWORD, {
+  'if', 'else', 'for', 'while', 'let', 'set', 'import', 'include', 'return',
+  'true', 'false', 'none', 'auto', 'not', 'in', 'and', 'or', 'as', 'show'
+})
+
+lex:set_word_list('UNITS', {'em', 'in', '%', 'mm', 'deg', 'rad', 'cm', 'pt', 'fr'})
+
+lex:add_fold_point(lexer.OPERATOR, '{', '}')
+lex:add_fold_point(lexer.COMMENT, '/*', '*/')
+lex:add_fold_point(lexer.PREPROCESSOR, '```', '```')
+lexer.property['scintillua.comment'] = '//'
+
+return lex
+