pratt-parser

Pratt Parser

Based on Top Down Operator Precedence and Douglas Crockford TDOP

import { Parser, WhiteSpaceToken, NumberToken } from "pratt-parser";

function Value(value) {
  return Object.create(null, {
    value: {
      value: value
    }
  });
}

const myGrammar = new Parser({
  tokens: [WhiteSpaceToken, NumberToken],
  prefix: {
    "(": {
      nud(grammar) {
        const e = grammar.expression(0);
        grammar.advance(")");
        return e;
      }
    }
  },
  infix: {
    ")": {},
    "+": {
      precedence: 50,
      combine: (left, right) => Value(left.value + right.value)
    },
    "-": {
      precedence: 50,
      combine: (left, right) => Value(left.value - right.value)
    },
    "*": {
      precedence: 60,
      combine: (left, right) => Value(left.value * right.value)
    },
    "/": {
      precedence: 60,
      combine: (left, right) => Value(left.value / right.value)
    }
  }
});

console.log(myGrammar.parse("(1 + (1 + 4 * 3)) * (2 + 1)").value);

API

Table of Contents

• ParsePosition
  • Properties
• Token
  • Properties
• RootToken
  • parseString
    • Parameters
• WhiteSpaceToken
• LineCommentToken
• EOFToken
• Parser
  • Parameters
  • error
    • Parameters
  • parse
    • Parameters
• Tokenizer
  • Parameters
  • tokens
    • Parameters
  • error
    • Parameters

ParsePosition

Type: Object

Properties

• offset number
• chunk string

Token

Type: Object

Properties

• type string
• precedence number
• value any

RootToken

Base object for all tokens

parseString

Parses from chunk of PrasePosition and delivers next token Modifies ParsePosition so that it points behind the detected token.

Parameters

• pp ParsePosition

Returns Token

WhiteSpaceToken

skip white space

LineCommentToken

skips until end of line

EOFToken

Token representing 'end of file'

Parser

Creates a grammar for later parsing

Parameters

• grammar any
• options Object

error

Forwards error to the tokenizer

Parameters

• args ...any

Returns Object error

parse

Parses the input and delivers the outermoost expression.

Parameters

• chunk string input text
• context Object object transparently passed to tokenizer

Returns Object evaluated input

Tokenizer

Creates a tokenizer for later parsing.

Parameters

• grammar Object definition of the grammar with operators...

tokens

delivers tokens from the input.

Parameters

• chunk string the input to be processed
• context Object additional info to be used by the actual token types

error

Parameters

• message string
• context Object token initiating the error
• value Object?

Returns Object error

install

With npm do:

npm install pratt-parser

license

BSD-2-Clause