|
| 1 | +local lpeg = require('lulpeg') |
| 2 | +local P, R, S, V = lpeg.P, lpeg.R, lpeg.S, lpeg.V |
| 3 | +local C = lpeg.C |
| 4 | + |
| 5 | +-- Some special symbols. |
| 6 | +local space_symbol = S(' \t\r\n') |
| 7 | +local eof = P(-1) |
| 8 | +local some_spaces = space_symbol ^ 0 |
| 9 | + |
| 10 | +-- |
| 11 | +-- Possible identifier patterns: |
| 12 | +-- 1) Number. |
| 13 | +local digit = R('09') |
| 14 | +local integer = (S('-+') ^ -1) * (digit ^ 1) |
| 15 | +local decimal = integer * '.' * (digit ^ 1) |
| 16 | +local number_pat = decimal + integer |
| 17 | +-- 2) Boolean. |
| 18 | +local bool_pat = P('false') + P('true') |
| 19 | +-- 3) String. |
| 20 | +local string_pat = P'\"' * C((P'\\"' + 1 - S'\"') ^ 0) * P'"' |
| 21 | +-- 4) Variable. |
| 22 | +local variable_name = ('_' + R('az', 'AZ')) * |
| 23 | + (('_' + R('09', 'az', 'AZ')) ^ 0) |
| 24 | +-- 5) Object's field path. |
| 25 | +local field = ('_' + R('az', 'AZ')) * ('_' + R('09', 'az', 'AZ')) ^ 0 |
| 26 | +local field_path = field * ('.' * field) ^ 0 |
| 27 | + |
| 28 | +-- |
| 29 | +-- Possible logical function patterns: |
| 30 | +-- 1) is_null. |
| 31 | +local is_null = P('is_null') |
| 32 | +-- 2) not_null. |
| 33 | +local not_null = P('not_null') |
| 34 | +-- 3) regexp. |
| 35 | +local regexp = P('regexp') |
| 36 | + |
| 37 | +-- |
| 38 | +-- Possible unary operator patterns: |
| 39 | +-- 1) Logical negation. |
| 40 | +local negation = P('!') |
| 41 | + |
| 42 | +-- Possible binary operator patterns: |
| 43 | +-- 1) Logical and. |
| 44 | +local logic_and = P('&&') |
| 45 | +-- 2) logical or. |
| 46 | +local logic_or = P('||') |
| 47 | +-- 3) + |
| 48 | +local addition = P('+') |
| 49 | +-- 4) - |
| 50 | +local subtraction = P('-') |
| 51 | +-- 5) == |
| 52 | +local eq = P('==') |
| 53 | +-- 6) != |
| 54 | +local not_eq = P('!=') |
| 55 | +-- 7) > |
| 56 | +local gt = P('>') |
| 57 | +-- 8) >= |
| 58 | +local ge = P('>=') |
| 59 | +-- 9) < |
| 60 | +local lt = P('<') |
| 61 | +-- 10) <= |
| 62 | +local le = P('<=') |
| 63 | + |
| 64 | +-- |
| 65 | +-- AST nodes generating functions. |
| 66 | +-- |
| 67 | +local function identical(arg) |
| 68 | + return arg |
| 69 | +end |
| 70 | + |
| 71 | +local function root_expr_node(expr) |
| 72 | + return { |
| 73 | + kind = 'root_expression', |
| 74 | + entity = expr |
| 75 | + } |
| 76 | +end |
| 77 | + |
| 78 | +local function bin_op_node(...) |
| 79 | + local args = {...} |
| 80 | + if select('#', ...) == 1 then |
| 81 | + return identical(args[1]) |
| 82 | + end |
| 83 | + local operators, nodes = {}, {} |
| 84 | + for i, v in ipairs(args) do |
| 85 | + if i % 2 == 0 then |
| 86 | + table.insert(operators, v) |
| 87 | + else |
| 88 | + table.insert(nodes, v) |
| 89 | + end |
| 90 | + end |
| 91 | + return { |
| 92 | + kind = 'binary_operations', |
| 93 | + operators = operators, |
| 94 | + nodes = nodes |
| 95 | + } |
| 96 | +-- return { |
| 97 | +-- kind = 'binary_operation', |
| 98 | +-- op = binary_operator, |
| 99 | +-- node_1 = operand_1, |
| 100 | +-- node_2 = operand_2 |
| 101 | +-- } |
| 102 | +end |
| 103 | + |
| 104 | +local function unary_op_node(unary_operator, operand_1) |
| 105 | + return { |
| 106 | + kind = 'unary_operation', |
| 107 | + op = unary_operator, |
| 108 | + node = operand_1 |
| 109 | + } |
| 110 | +end |
| 111 | + |
| 112 | +local function func_node(name, ...) |
| 113 | + local args = {...} |
| 114 | + return { |
| 115 | + kind = 'function', |
| 116 | + name = name, |
| 117 | + args = args |
| 118 | + } |
| 119 | +end |
| 120 | + |
| 121 | +local function op_name(string) |
| 122 | + return string |
| 123 | +end |
| 124 | + |
| 125 | +local function number_node(value) |
| 126 | + return { |
| 127 | + kind = 'const', |
| 128 | + data_class = 'number', |
| 129 | + value = value |
| 130 | + } |
| 131 | +end |
| 132 | + |
| 133 | +local function string_node(value) |
| 134 | + return { |
| 135 | + kind = 'const', |
| 136 | + data_class = 'string', |
| 137 | + value = value |
| 138 | + } |
| 139 | +end |
| 140 | + |
| 141 | +local function bool_node(value) |
| 142 | + return { |
| 143 | + kind = 'const', |
| 144 | + data_class = 'bool', |
| 145 | + value = value |
| 146 | + } |
| 147 | +end |
| 148 | + |
| 149 | +local function variable_node(name) |
| 150 | + return { |
| 151 | + kind = 'variable', |
| 152 | + name = name |
| 153 | + } |
| 154 | +end |
| 155 | + |
| 156 | +local function path_node(path) |
| 157 | + return { |
| 158 | + kind = 'object_field', |
| 159 | + path = path |
| 160 | + } |
| 161 | +end |
| 162 | + |
| 163 | +-- Patterns returning corresponding nodes. |
| 164 | +local _number = number_pat / number_node |
| 165 | +local _bool = bool_pat / bool_node |
| 166 | +local _string = string_pat / string_node |
| 167 | +local _variable = '$' * C(variable_name) / variable_node |
| 168 | +local _field_path = field_path / path_node |
| 169 | + |
| 170 | +local _logic_or = logic_or / op_name |
| 171 | +local _logic_and = logic_and / op_name |
| 172 | +local _comparison_op = (eq + not_eq + ge + gt + le + lt) / op_name |
| 173 | +local _arithmetic_op = (addition + subtraction) / op_name |
| 174 | +local _unary_op = negation / op_name |
| 175 | +local _functions = (is_null + not_null + regexp) / identical |
| 176 | + |
| 177 | +-- |
| 178 | +-- Grammar rules for C-style expressions positioned ascending in |
| 179 | +-- terms of priority. |
| 180 | +-- |
| 181 | +local expression_grammar = P { |
| 182 | + 'init_expr', |
| 183 | + init_expr = V('expr') * eof / root_expr_node, |
| 184 | + expr = some_spaces * V('log_expr_or') * some_spaces / identical, |
| 185 | + |
| 186 | + log_expr_or = V('log_expr_and') * (some_spaces * _logic_or * |
| 187 | + some_spaces * V('log_expr_and')) ^ 0 / bin_op_node, |
| 188 | + log_expr_and = V('comparison') * (some_spaces * _logic_and * |
| 189 | + some_spaces * V('comparison')) ^ 0 / bin_op_node, |
| 190 | + comparison = V('arithmetic_expr') * (some_spaces * _comparison_op * |
| 191 | + some_spaces * V('arithmetic_expr')) ^ 0 / bin_op_node, |
| 192 | + arithmetic_expr = V('unary_expr') * (some_spaces * _arithmetic_op * |
| 193 | + some_spaces * V('unary_expr')) ^ 0 / bin_op_node, |
| 194 | + |
| 195 | + unary_expr = (_unary_op * V('first_prio') / unary_op_node) + |
| 196 | + (V('first_prio') / identical), |
| 197 | + first_prio = (V('_function') + V('identifier') + '(' * some_spaces * |
| 198 | + V('expr') * some_spaces * ')') / identical, |
| 199 | + _function = _functions * '(' * some_spaces * (V('identifier') * |
| 200 | + some_spaces * ',' * some_spaces) ^ 0 * V('identifier') |
| 201 | + * some_spaces * ')' / func_node, |
| 202 | + identifier = (_bool + _number + _string + _variable + _field_path) / |
| 203 | + identical |
| 204 | +} |
| 205 | + |
| 206 | +-- |
| 207 | +-- |
| 208 | +-- |
| 209 | + |
| 210 | + |
| 211 | +-- |
| 212 | +-- Parsing function. |
| 213 | +-- |
| 214 | +return function(str) |
| 215 | + assert(type(str) == 'string', 'parser expects a string') |
| 216 | + return expression_grammar:match(str) or error('syntax error', 2) |
| 217 | +end |
0 commit comments