Node Fundamentals: Modules

This is a series of posts that will illustrate the what, why and how of Node. I'll be sharing my learnings from a course on Advanced NodeJS by Samer Buna offered on PluralSight. Any code samples tagged or attached will be available at the following repo.

Node Modules

Modularity is a first-class concept in Node.

There are two core modules involved.

1. require - It is a global function, but each module gets its own require function

2. module - It is also available globally and is used to manage all the modules we require with require.

Requiring a module in node is very simple concept.

To execute a require call, node goes through a following sequence of steps:

1. Resolving: Find the absolute file path of the module required.
2. Loading: Determined by the content of the file at the resolved path.
3. Wrapping: Gives every module its private scope and what makes require local to every module.
4. Evaluating: Eventually, VM does something to code.
5. Caching: When we require again, we don't go over all steps above mentioned.

module Object

Some interesting properties:

• id: String Identifier, Usually full path to the module except for the root module. . identifier is used for the root module.
• filename: String path to the file containing the module. So when you require a module from node_modules, it loads the content of a file into memory.
• path: Array of paths that will used to find a module that is required. It starts with node_modules folder in current directory and goes all the way to the root directory. If it can't find a module in any of those directories, it will throw a Cannot find module 'module' error. Core node modules are an exception. When you require a core node module, it resolves immediately.

Let's consider following example.

// ./index.js
console.log("In ./index.js");
require("find-me");

// ./node_modules/find-me.js
console.log("In find-me.js");

This will result in output

In ./index.js
In find-me.js

Note, not only it loads the file, it also evaluates when you require it.

If you only want to load a file, and not evaluate it, you can use require.resolve(package). It will also throw an error if it can't find the package in any of the paths. This is used to determine if an optional package is installed or not.

If you have package installed in multiple folders present in path property, it will only resolve the first one it finds.

Usually, packages are not files, but are folders, with multiple files. It will use index.js or main property in package.json of the package. We can require any module, with a relative or absolute path.

Note module object available in index.js and in package index.js are different. module object in package index.js will have a reference to root index.js, and will be attached to it's parent property.

Module Wrapper

index.js

exports.id = 1 // this is ok
exports = { id: 1 } // this is not ok

module.exports = { id: 1 } // this is ok, why?

var  = 42; // local to this file

Only the things we export are available outside the module. How come variables we declare are magically scope. The answer is simple.

Before compiling a module, Node will wraps the module code in a function as follows.

> require('module').wrapper

(function (exports, require, module, __filename, __dirname) { ',
    '\n});

This is how each module gets its own require, exports and module object. These are just function arguments that are provided by wrapped function by node.

To see the values of these arguments you can just run the following code.

// index.js

console.log(arguments);

This will print all 5 arguments passed to the wrapper function.

The wrapping function's return value is the exports object reference. Note, exports is just a variable reference to module.exports. So, if we modify the whole of exports by assignment operator, we lose the module.exports module.

So, there's nothing special about require function. It takes the module name or path and returns the exports object. So in test case scenarios, where one might need to overwrite/mock require, we can do quite easily as follows.

require = () => {
  return { mocked: true };
};

console.log(require("somepackage")) // { mocked: true }

Let's say we have this simple function that takes an integer and a string and prints something.

// printStars.js
const print = (stars, header) => {
  console.log("*".repeat(stars));
  console.log(header);
  console.log("*".repeat(stars));
}

We want to run this function in two ways.

1. Through the command line as follows

$ node printStars.js 5 hello

1. Through require in another file as a module as follows.

// index.js

const printStars = require("printStars");
print(5, "hello");

To achieve this, we can leverage wrapping.
When it is run through Node CLI, require.main will be the same as module.

//printStars.js

const print = (stars, header) => {
  console.log("*".repeat(stars));
  console.log(header);
  console.log("*".repeat(stars));
}

if (require.main == module) {
  // When run as script
  print(process.argv[2], process.argv[3]);
} else {
  // being required by other file
  // export the module
  module.exports = print;
}

Caching

Imagine this case

// index.js

require("printFancy"); // Prints
console.log(require.cache); // Entry for `printFancy` module
require("printFancy"); // Nothing happens 

//printFancy.js

console.log("Hello Fancy");

Note, when we require printFancy first time, it will resolve, load, evaluate and cache the module.

However, when we require again, node has cached the module and so will repeat earlier steps again.

To circumvent this we can delete the cache on require.cache object, before the second call, with delete require.cache['absModulePath'] and printFancy will be called twice. But it is not the most efficient solution.

The easiest solution is to wrap the console log in printFancy.js in a function and export it.

// printFancy.js
module.exports = () => {
  console.log("Hello Fancy");
};

Now every time you require the module, just execute the exports.

// index.js

require('printFancy')() // Prints 
require('printFancy')() // Prints