CodeTree

CodeTree is an ES6 library (written in TypeScript) that helps transform code easily.

Transforming TypeScript or JavaScript code

import * as codeTree from "codetree";
 
// This is the JavaScript (or TypeScript) code we want to transform.
let code = `
  function sayHelloWorld() {
    console.log("Hello, World!");
  }
`;
 
// Convert the source code to a parse tree.
let tree = codeTree.typescript.parse(code);
 
// Create a mutated tree.
let mutatedTree = codeTree.transformTree(
  tree,
  // Match any console.log() statement. Double underscore means "anything".
  codeTree.typescript.parse("console.log(__);"),
  // Replace with console.error("Hi!");
  matchedTree => codeTree.typescript.parse(`console.error("Hi!");`)
);
 
// Convert the mutated tree back into source code.
// Your comments and formatting will not be lost. Yay!
console.log(codeTree.printSource(mutatedTree));
 

Give it a try yourself here: https://runkit.com/fwouts/using-codetree-to-mutate-javascript

Transforming any other language

Any language that can be parsed with ANTLR4 can be transformed with CodeTree. That is, virtually any language since open-source contributors have been kind enough to provide an extensive library of grammars in the grammars-v4 repo.

Creating the parser and lexer

After downloading the ANTLR grammar for your language (one combined or two parser + lexer files ending with the .g4 extension), you'll need to generate the TypeScript lexer and parser before you can use CodeTree.

Here is an example using the Java grammar:

$ npm install -g antlr4ts-cli
$ antlr4ts -no-listener -no-visitor *.g4
Generating file '/Users/work/GitHub/abc/./JavaLexer.ts' for grammar 'JavaLexer.g4'
Generating file '/Users/work/GitHub/abc/./JavaLexer.tokens' for grammar 'JavaLexer.g4'
Generating file '/Users/work/GitHub/abc/./JavaParser.ts' for grammar 'JavaParser.g4'
Generating file '/Users/work/GitHub/abc/./JavaParser.tokens' for grammar 'JavaParser.g4'

This will generate two new files: JavaLexer.ts and JavaParser.ts. You can ignore the .token files and even delete them. They're not necessary.

If you work with JavaScript and not TypeScript, convert these files to ES6 using:

$ npm install -g typescript
$ npm install antlr4ts
$ tsc *.ts --target es6 --experimentalDecorators --moduleResolution node
$ mv *.js path/to/src/

Using the parser and lexer

Now, all you need to do is:

import * as codeTree from "codetree";
 
import { JavaLexer } from "./parsers/java/JavaLexer";
import { JavaParser } from "./parsers/java/JavaParser";
 
// This is the Java code we want to transform.
let code = `
  package example;
 
  public class Example {
    public static void main() {
      System.out.println("Hello, World!");
    }
  }
`;
 
// Convert the source code to a parse tree.
let tree = codeTree.antlr.parse(
  JavaLexer,
  JavaParser,
  // This tells CodeTree that we need a CompilationUnit node. This will depend on the
  // grammar you use.
  parser => parser.compilationUnit(),
  code
);
// Just making sure we got a tree and not a terminal node.
if (!(tree instanceof codeTree.nodes.Tree)) {
  throw new Error();
}
 
// Create the mutated tree.
let mutatedTree = codeTree.transformTree(
  tree,
  codeTree.antlr.parse(
    JavaLexer,
    JavaParser,
    // This tells CodeTree that we're looking to match a statement.
    parser => parser.statement(),
    // Match any statement that is a call to System.out.println().
    `System.out.println(__);`
  ),
  matchedTree =>
    codeTree.antlr.parse(
      JavaLexer,
      JavaParser,
      // This tells CodeTree that we're replacing with a statement.
      parser => parser.statement(),
      // Replace with alert("Hi!");
      `alert("Hi!");`
    )
);
 
// Convert the mutated tree back into source code.
console.log(codeTree.printSource(mutatedTree));