Creates an AST parser from a [E]BNF file
npm install ebnf
It parses a formal grammar and returns a parser instance. That parser returns ASTs based on your grammar. SEE IT IN ACTION HERE GitHub source
npm i --save ebnf (It's compatible with WebPack, and Browserify)
By the moment we only accept two grammars. BNF and W3C EBNF (compatible with Railroad Diagram Generator)
``typescript
import { Grammars } from 'ebnf';
let bnfParser = new Grammars.BNF.Parser(bnfGrammar);
let w3cParser = new Grammars.W3C.Parser(railRoadGeneratorGrammar);
`
Check out the test folder for more examples
In this example we use plain BNF to create a simple integer formula parser
Grammar:
`ebnf
`
`typescript
import { Grammars } from 'ebnf';
let parser = new Grammars.BNF.Parser(grammar);
parser.getAST('-122 + 2');
/* -122 + 2 Equation
-122 + 2 BinaryOperation
-122 Term
-122 Number
+ Operator
2 Term
2 Number
*/
parser.getAST( '(2 + (2 -123)) 5332');
/ (2 + (2 -123)) * 5332 Equation
(2 + (2 -123)) 5332 BinaryOperation
(2 + (2 * -123)) Term
2 + (2 * -123) Equation
2 + (2 * -123) BinaryOperation
2 Term
2 Number
+ Operator
(2 * -123) Term
2 * -123 Equation
2 * -123 BinaryOperation
2 Term
2 Number
* Operator
-123 Term
-123 Number
* Operator
5332 Term
5332 Number
*/
`
`wbnf
/ https://www.ietf.org/rfc/rfc4627.txt /
value ::= false | null | true | object | array | number | string
BEGIN_ARRAY ::= WS #x5B WS / [ left square bracket /
BEGIN_OBJECT ::= WS #x7B WS / { left curly bracket /
END_ARRAY ::= WS #x5D WS / ] right square bracket /
END_OBJECT ::= WS #x7D WS / } right curly bracket /
NAME_SEPARATOR ::= WS #x3A WS / : colon /
VALUE_SEPARATOR ::= WS #x2C WS / , comma /
WS ::= [#x20#x09#x0A#x0D]+ / Space | Tab | \n | \r /
false ::= "false"
null ::= "null"
true ::= "true"
object ::= BEGIN_OBJECT (member (VALUE_SEPARATOR member)*)? END_OBJECT
member ::= string NAME_SEPARATOR value
array ::= BEGIN_ARRAY (value (VALUE_SEPARATOR value)*)? END_ARRAY
number ::= "-"? ("0" | [1-9] [0-9]) ("." [0-9]+)? (("e" | "E") ( "-" | "+" )? ("0" | [1-9] [0-9]))?
/ STRINGS /
string ::= '"' (([#x20-#x21] | [#x23-#x5B] | [#x5D-#xFFFF]) | #x5C (#x22 | #x5C | #x2F | #x62 | #x66 | #x6E | #x72 | #x74 | #x75 HEXDIG HEXDIG HEXDIG HEXDIG))* '"'
HEXDIG ::= [a-fA-F0-9]
`
`typescript
import { Grammars } from 'ebnf';
let parser = new Grammars.W3C.Parser(grammar);
parser.getAST( '{"a":false,"b":"asd\\n asd ","list":[1,2,3,true]}');
/* {"a":false,"b":"asd\n asd ","list":[1,2,3,true]} value
{"a":false,"b":"asd\n asd ","list":[1,2,3,true]} object
"a":false member
"a" string
false value
false false
"b":"asd\n asd " member
"b" string
"asd\n asd " value
"asd\n asd " string
"list":[1,2,3,true] member
"list" string
[1,2,3,true] value
[1,2,3,true] array
1 value
1 number
2 value
2 number
3 value
3 number
true value
true true
*/
`
Every ast node has the following interface
`typescript`
interface IToken {
type: string; // Rule name
text: string; // Inner text
children: IToken[]; // Children nodes
start: number; // Start position of the input string
end: number; // End position
errors: TokenError[]; // List of Errors
}
We try to keep this tool as much unopinionated and free of conventions as possible. However, we have some conventions:
- All UPPER_AND_SNAKE_CASE rules are not emmited on the AST. This option can be deactivated setting the flag keepUpperRules: true`.