@babel/generator
Turns Babel AST into code.
Install
- npm
- Yarn
- pnpm
npm install --save-dev @babel/generator
yarn add --dev @babel/generator
pnpm add --save-dev @babel/generator
Usage
import { parse } from "@babel/parser";
import generate from "@babel/generator";
const code = "class Example {}";
const ast = parse(code);
const output = generate(
ast,
{
/* options */
},
code
);
The symbols like white spaces or new line characters are not preserved in the AST. When Babel generator prints code from the AST, the output format is not guaranteed.
Parser plugins support
Babel generator supports all the listed Babel parser plugins except estree
. Note that parser plugins do not transform the code. For example,
if you pass JSX <div></div>
to babel generator, the result will still contain the div
JSX element.
import { parse } from "@babel/parser";
import generate from "@babel/generator";
const code = "const Example = () => <div>example</div>";
const ast = parse(code, { plugins: ["jsx" ] });
const output = generate(
ast,
);
// true
output.includes("<div>");
Options
History
Version | Changes |
---|---|
v7.26.0 | Added experimental_preserveFormat |
v7.22.0 | Added importAttributesKeyword |
v7.21.0 | Added inputSourceMap |
Options for formatting output:
name | type | default | description |
---|---|---|---|
auxiliaryCommentAfter | string | Optional string to add as a block comment at the end of the output file | |
auxiliaryCommentBefore | string | Optional string to add as a block comment at the start of the output file | |
comments | boolean | true | Should comments be included in output |
compact | boolean or 'auto' | opts.minified | Set to true to avoid adding whitespace for formatting |
concise | boolean | false | Set to true to reduce whitespace (but not as much as opts.compact ) |
decoratorsBeforeExport | boolean | Set to true to print decorators before export in output. | |
filename | string | Used in warning messages | |
importAttributesKeyword | "with" , "assert" , or "with-legacy" | The import attributes/assertions syntax to use. "with" for import "..." with { type: "json" } , "assert" for import "..." assert { type: "json" } , and "with-legacy" for import "..." with type: "json" . When not specified, @babel/generator will try to match the style in the input code based on the AST shape. | |
jsescOption | object | Use jsesc to process literals. jsesc is applied to numbers only if jsescOption.numbers (added in v7.9.0 ) is present. You can customize jsesc by passing options to it. | |
jsonCompatibleStrings | boolean | false | Set to true to run jsesc with "json": true to print "\u00A9" vs. "©"; |
minified | boolean | false | Should the output be minified |
retainFunctionParens | boolean | false | Retain parens around function expressions (could be used to change engine parsing behavior) |
retainLines | boolean | false | Attempt to use the same line numbers in the output code as in the source code (helps preserve stack traces) |
shouldPrintComment | function | opts.comments | Function that takes a comment (as a string) and returns true if the comment should be included in the output. By default, comments are included if opts.comments is true or if opts.minified is false and the comment contains @preserve or @license |
recordAndTupleSyntaxType | 'hash' or 'bar' | 'hash' | For use with the recordAndTuple token. |
topicToken | '%' or '#' | The token to use with Hack-pipe topic references. This is required when there are any TopicReference nodes. |
Options for source maps:
name | type | default | description |
---|---|---|---|
sourceMaps | boolean | false | Enable generating source maps |
inputSourceMap | string or object | The input source map | |
sourceRoot | string | A root for all relative URLs in the source map | |
sourceFileName | string | The filename for the source code (i.e. the code in the code argument). This will only be used if code is a string. |
Experimental options:
The behavior of experimental options could have breaking changes in minor versions.
name | type | default | description | experimental reason |
---|---|---|---|---|
experimental_preserveFormat | boolean | false | When set to true , the generator will try to preserve location of all the nodes and tokens that are present both in the input and output code. To use this option, you currently need to enable the retainLines: true generator option, and the tokens: true and createParenthesizedExpressions: true parser options. | This option will graduate to stable once it supports a mode that does not require preserving line numbers, but just relative positions of tokens. |
AST from Multiple Sources
In most cases, Babel does a 1:1 transformation of input-file to output-file. However,
you may be dealing with AST constructed from multiple sources - JS files, templates, etc.
If this is the case, and you want the sourcemaps to reflect the correct sources, you'll need
to pass an object to generate
as the code
parameter. Keys
should be the source filenames, and values should be the source content.
Here's an example of what that might look like:
import { parse } from "@babel/parser";
import generate from "@babel/generator";
const a = "var a = 1;";
const b = "var b = 2;";
const astA = parse(a, { sourceFilename: "a.js" });
const astB = parse(b, { sourceFilename: "b.js" });
const ast = {
type: "Program",
body: [].concat(astA.program.body, astB.program.body),
};
const { code, map } = generate(
ast,
{ sourceMaps: true },
{
"a.js": a,
"b.js": b,
}
);
// Sourcemap will point to both a.js and b.js where appropriate.