kaelzhang / neuron.js

:shipit: A Full Feature CommonJS Module Manager, Dependency Graph Handler and Loader for Browsers

Neuron

First of all, neuron is not designed for human developers to use directly. Most usually, it works together with neuron-cli.

Neuron is a full feature CommonJS module loader which makes your node-style modules run in browsers.

• Dynamic resource loading with "module -> resource url" mapping.
• Implements commonjs Module/1.0 standard.
• Fully supports SemVer and SemVer ranges: '^a.b.c', '~a.b.c', '>=a.b.c', etc.
• Implements File Modules of node.js (Maybe the only module loader which could do that.)
• Supports cyclic dependencies.
• Implements require.resolve(), __filename, and __dirname for browsers which is similar to node.js.
• Completely isolated sandboxes.
• Supports scoped packages

With Neuron, we write web modules exactly the same as we work with node.js, with no Module/Wrappings, no *MD, etc.

Neuron is designed to run in the background without your concern, UNLIKE RequireJS and many other loaders.

---

Neuron Loader for Browsers

Frequent configurations, for more, just see Configuration section.

<script src="/dist/neuron.js"></script>
<script>
facade('hello', {
  name: 'John'
});
</script>

Install and build dist

npm install
node node/build

With ecma5 compatibility

npm install
node node/build ecma5

facade()

facade(identifier);
facade(identifier, data);

• identifier String module name with version, seperated with '@'. For example: 'async@0.1.0'

• data Object will be passed as the parameter of the module.exports.

Method facade loads a module. If the module.exports is a function, facade method will run the function with data as its only parameter.

We call this kind of modules as facade modules which is much like the bins of nodejs.

require(id)

• id String module identifier.

To require modules. See CommonJS Module/1.0

require.async(id, callback)

• id String module identifier.
• callback function(exports) callback must be passed, or require.async will do nothing.

Asynchronously loads a module by id, and then passes the module exports to callback.

You should always pass the callback parameter because neuron can not make sure the exact time when a module is loaded asynchronously.

It is NOT a good practice if the logic of your code relies on the result of the require.async()d module without a callback.

require.resolve(path)

• path String the relative path to be resolved according to the current module.

Returns the resolved absolute path of the resource.

Returns undefined if path is not a relative path.

Returns undefined if path is even outside the current package.

---

Developer Guide

Neuron CORE supplies no high-level APIs, which means that neuron core only cares about module dependencies and module wrapping while will do nothing about things such as fetching modules from remote server and injecting them into the current document, and never cares about where a specific module should come from.

Configuration

neuron.config(settings);

settings.resolve function(id)

Method to resolve the module id into url paths.

By default, it works with settings.path, and resolves the module id into

settings.path
+ id
  // '@facebook/react@1.0.0/react.js' -> 'facebook/react@1.0.0/react.js'
  .replace(/^@/, '')
  // 'facebook/react@1.0.0/react.js'
  // -> 'facebook/react/1.0.0/react.js'
  // -> '/mod/facebook/react@1.0.0/react.js'
  .replace('@', '/');

settings.path String

CommonJS module path, like NODE_PATH, default to '/mod/'.

Pay attension that path will not be resolved to absolute url. So if you don't want a relative path, don't forget 'http://'.

Actually, settings.path only works with settings.resolve and provides a simple way to customize the base path of the package resources. If you defines your own settings.resolve, settings.path will be useless.

settings.loaded String|Array.<id>

To tell neuron loader that those modules are already loaded, and prevent duplicate loading.

If String, we can separate different ids with '|' (comma).

neuron.config({
	loaded: ['jquery@1.9.2', 'async@0.2.9']
});

settings.graph Object

The directed graph of all dependencies, which could be parsed by neuron-graph.

The arithmetics to generate the graph is complicated and hard to describe, see https://github.com/kaelzhang/neuron/blob/master/doc/graph.md for details (Too Long; Don't Read)

define()

You should NEVER write this method by hands.

ALWAYS use builders(such as neuron-builder) to generate this method.

define(identifier, dependencies, factory, options);

id (full module id)

Format: <package-name>@<version>/<path-with-extension>

Type: string

The real pathname relative to the root directory of the package.

dependencies

Array.<id>

factory

function(require, exports, module, __filename, __dirname){}

options

options.main

Type Boolean

whether the module is the main entry, i.e. the package.main field in package.json

options.map

Type Object

<id>:<full-module-id>.

require('./a')
require('./lib')
// ->
// map: {
//   './a': 'my@1.0.0/a.js'
//   // require a directory
//   './lib': 'my@1.0.0/lib/index.js'
// }

---

NPM module: neuron.js

A package to get the JavaScript file of neuron.

var neuron = require('neuron.js');
neuron.version(); // 10.1.0
neuron.content(function(err, content){
  content; // The file content of neuron.js
});

neuron.version();

Returns String the version of neuron for browsers, not the version of npm module neuronjs

neuron.write(dest, callback)

• dest path
• callback function(err)

Writes the content of neuron.js to the dest

neuron.content(callback)

• callback function(err, content)
• content Buffer the buffer of the content of neuron.js

Gets the content of neuron.js

Related Projects

• cortex
• neocortex-sync