eslint-traverser

A utility that helps traverse code the way ESLint does. This allows you to test any utilities you write for ESLint rules.

Installation

npm install eslint-traverser

Usage

Common Usage

The module exposes a function that gets code and an optional config object, and gets a traverser, with a get function. The get function calls the callback for every node of the type, with node and context parameters, which are the same as the ones in ESLint itself.

const traverse = require('eslint-traverser')

traverse('var y = f(x)')
    .get('CallExpression', (node, context) => {
      console.log(node.callee.name) //logs `f`
      sourceCode = context.getSourceCode()
      //...
    })

Using configs

You can define a configuration to run your traversal (globals, settings, etc) in an additional parameter in the call to traverser. You may not use the rules key in this configuration.

const traverse = require('eslint-traverser')

traverse('import foo from "bar"', {parserOptions: {sourceType: module}})
    .get('Program:exit', (node, context) => {
      console.log('Modules!')
    })

As a shortcut, you can pass only the parserOptions object, and if your config only contains parserOptions keys, it will work normally.

const traverse = require('eslint-traverser')

traverse('import foo from "bar"', {sourceType: module})
    .get('Program:exit', (node, context) => {
      console.log('Modules!')
    })

Using a selector

You can also filter results using a selector, which can be a function or any of the iteratees supplied by Lodash

const traverse = require('eslint-traverser')

traverse('f(); g(x)')
    .get('CallExpression', node => node.arguments.length, (node, context) => {
      console.log(node.callee.name) //logs `g`
      //...
    })
    
traverse('f(); g(x)')
    .get('CallExpression', 'arguments.length', node => {
      console.log(node.callee.name) //logs `g`
      //...
    })
    
traverse('f(); g(x)')
    .get('CallExpression', ['arguments.length', 0], node => {
      console.log(node.callee.name) //logs `f`
      //...
    })
    
traverse('f(); g(x)')
    .get('CallExpression', {callee: {name: 'g'}}, node => {
      console.log(node.callee.name) //logs `g`
      //...
    })    

Using the first method

The first method is the same as the get method, except it only calls the callback on the first result, while get calls it on all the results

const traverse = require('eslint-traverser')
traverse('f(); g()')
    .get('CallExpression', node => {
      console.log(node.callee.name) 
    })
// f
// g

traverse('f(); g()')
    .first('CallExpression', node => {
      console.log(node.callee.name)
    })
// f

Using the visitAll method

The visitAll method gets two arguments: first, an optional visitors object in the same structure as an ESLint rule visitor object. The second is a callback, that gets called at the end of the code traversal with two arguments: the Program node and the context.

const traverse = require('eslint-traverser')
traverse('var y = f(x)')
    .visitAll((node, context) => {
      console.log(node.type) //Program
      
    })

traverse('var y = f(x)')
    .visitAll({
      CallExpression(node) { console.log('Call expression!')}
    }, () => { console.log('finished!')})
// Call expression!
// finished!

Using the runRuleCode method

The runRuleCode method gets one argument: the rule. This runs an ESLint rule (a function that gets a context and returns a visitors object) on the specified code.

const traverse = require('eslint-traverser')
traverse('var y = f(x)')
    .runRuleCode(context => ({
      CallExpression(node) { console.log('Call expression!')}
    }))
// Call expression!