amd-utils

http://millermedeiros.github.com/amd-utils

Modular JavaScript utilities written in the AMD format.

All code is library agnostic and consist mostly of helper methods that aren't directly related with the DOM, the purpose of this library isn't to replace Dojo, jQuery, YUI, Mootools, etc, but to provide modular solutions for common problems that aren't solved by most of them.

Main goals

increase code reuse;
be clear (code should be clean/readable);
be easy to debug;
be easy to maintain;
follow best practices;
follow standards when possible;
don't convert JavaScript into another language!
be compatible with other frameworks;
be modular;
have unit tests for all modules.

What shouldn't be here

UI components;
CSS selector engine;
Event system - pub/sub;
Template engine;
Anything that could be a separate library and/or isn't a modular utility...

Why AMD?

Because AMD is awesome!

By keeping each function in a separate package we can require just the methods/packages that are required by our app and let the RequireJS optimizer bundle only what is currently being used. We also have an extra benefit that we split the methods into separate packages so we reduce the chance of name collisions and the code is more organized. AMD is flexible and enables things that wouldn't be possible with a different module system or with a conventional namespace approach (remapping a module, conditionally loading, etc).

Read these links if you still can't see why:

PS: Your mileage may vary...

Node.js

AMD-utils also works on node.js, just run:

npm install amd-utils

It will download amd-utils from the NPM repository and convert the AMD modules into a node.js compatible format using nodefy, there is no extra overhead, you can use it like a regular node.js package.

// you can load individual methods
var map = require('amd-utils/array/map');
map([1, 2], function(v){ return val * val; }); // [1, 4]

// a single package
var stringUtils = require('amd-utils/string');
stringUtils.camelCase('Foo Bar'); // "fooBar"

// or the whole lib
var utils = require('amd-utils');
console.log( utils.math.clamp(100, 0, 50) ); // 50

Important

Since each function is a separate module they have independent version numbers and a last edit date.

Since code is very modular (broken into multiple files) it is really important that you run an optimizer before deploying the code to a server, otherwise you may end up having too many file requests which can degrade load-time performance a lot. See RequireJS optimization for more info.

Contributing

Fork the project on Github: https://github.com/millermedeiros/amd-utils

"Write clearly, don't be too clever" - The Elements of Programming Style

Avoid unnamed functions and follow the other modules structure. By only using named functions it will be easier to extract the code from the AMD module if needed and it will also give better error messages, JavaScript minifiers like Google Closure Compiler and UglifyJS will make sure code is as small/optimized as possible.

"Make it clear before you make it faster." - The Elements of Programming Style

Be sure to always create tests for each proposed module. Features will only be merged if they contain proper tests and documentation.

"Good code is its own best documentation." - Steve McConnell

We should do a code review before merging to make sure names makes sense and implementation is as good as possible.

Check the contributors list at github.

Build Script

The build script have a set of very helpful commands, run npm install --dev (only required once) and then check the available commands:

node build --help

Keeping packages and specs in sync

The build script can be used to update packages and specs files:

node build pkg

The packages/specs are automatically updated every time you run npm test as well.

You can also add new modules with the command node build add package/moduleName, this will create a new module src/package/moduleName.js and also a failing spec tests/spec/package/spec-moduleName.js.

Tests & Code Coverage

Tests can be found inside the tests folder, to execute them in the browser open the tests/runner.html. The same tests also work on node.js by running npm test.

We should have tests for all methods and ensure we have a high code coverage through our continuous integration server (travis). When you ask for a pull request Travis will automatically run the tests on node.js and check the code coverage as well.

We run node build pkg automatically before any npm test, so specs and packages should always be in sync. (will avoid human mistakes)

To check code coverage run npm test --coverage, it will generate the reports inside the coverage folder and also log the results. Please note that node.js doesn't execute all code branches since we have some conditionals that are only met on old JavaScript engines (eg. IE 7-8), so we will never have 100% code coverage (but should be close to it).

Building The Documentation

The documentation is generated based on markdown files inside the doc/mdown folder using mdoc. To compile the docs run:

node build doc

It will replace all the files inside the doc/html folder.

Documentation files should be always up-to-date since modules are only committed to the master branch after they get proper tests and documentation.

Documentation

Online documentation can be found at http://millermedeiros.github.com/amd-utils or inside the doc folder.