Config files that defer to env settings.


Licenses
MIT/MIT
Install
npm install configya@0.2.3

Documentation

configya

Stupid simple configuration.

What & How

configya reads the following:

  • environment variables
  • optional configuration
  • defaults hash

Note: you can now provide both a configuration file and a defaults object

Unless you've set a deploy-type environment variable = 'DEV', configya will always overwrite keys from a configuration file and defaults hash with duplicates found in the environment.

Key Parsing

configya parses all sources into an object hierarchy based on _ delimited key names (If camelCase support is desired, see below). For example, if you have a key named RABBIT_BROKER_IP set to '127.0.0.1', and another named RABBIT_BROKER_PORT set to 5672, the resulting configuration object will be:

{
	rabbit: {
		broker: {
			ip: "127.0.0.1",
			port: "5672"
		}
	}
}

Note: When using a single _ delimeter, all keys are lower-cased to eliminate the need for guessing games (and capslock)

Camel case is supported when configya detects a double underscore __ in the key. Once found, _ will be treated as a word boundary instead of a nesting separator. If you have a key named SQL__USER_NAME (notice the single underscore _ between USER and NAME) set to 'siteuser', and another named TARGET_SERVER__ set to 'localhost', the resulting configuration object will be:

{
    sql: {
        userName: "siteuser"
    },
    targetServer: "localhost"
}

In order to support top level keys with camelCase, you'll need to use a prefix, or start or end your key with __ (as shown above with TARGET_SERVER__).

Key Prefixing

configya lets you specify a prefix for your environment variables to be removed when your configuration is created. For example, if you have an environment variable LK_RABBIT_BROKER_PORT set to 5672, and call configya with the prefix option var cfg = require( 'configya' )({prefix:'lk'});, the resulting configuration object will be:

{
    rabbit: {
        broker: {
            port: "5672"
        }
    }
}

Note: If you are overriding a key that needs camelCase support, be sure to separate your prefix with __ intead of _: LK__CAMEL_CASE.

Value Parsing

Under the hood, configya uses JSON.parse to attempt to parse the values so "true" and "false" will become actual booleans, and "123" will be a number.

The following enviroment variables HOST_CACHE=false HOST_PORT=1234 FONTSIZE=10p would be parsed as follows:

{
    host: {
        cache: false,
        port: 1234
    },
    fontsize: "10p"
}

This means you can also use JSON as the value for an environment variable and it will be correctly parsed into an object.

Original Keys

The original keys are technically still stored on the object based on their source.

  • _env_ for original environment keys
  • _defaults_ for keys coming from a defaults hash
  • _file_ for keys coming from a file

Note: These are really here for diagnostic/backwards compatibility. You shouldn't use/rely on them in your code.

Usage

	//without a config file or defaults (using only environment)
	var cfg = require( 'configya' )();

    //with an environment prefix
    var cfg = require( 'configya' )({prefix: 'lk'});

	//with a config file as well
	var cfg = require( 'configya' )({file: './path/to/configuration.json'});

	//with a defaults hash
	var cfg = require( 'configya' )({
        defaults: { RABBIT_BROKER_PORT: 5672 }
    });

	//with defaults and a config
	var cfg = require( 'configya' )({
        defaults:{
            rabbit: { broker: { port: 5672 } }
        },
        file: './path/to/configuration.json'
    });

	var port = cfg.rabbit.broker.port; // etc.

Backwards Compatibility

The previous version of configya accepted multiple arguments of either a string for the file path or an object literal for defaults.

    //without a config file or defaults (using only environment)
    var cfg = require( 'configya' )();

    //with a config file as well
    var cfg = require( 'configya' )( './path/to/configuration.json' );

    //with a defaults hash
    var cfg = require( 'configya' )( { RABBIT_BROKER_PORT: 5672 } );

    //with defaults and a config (order of args doesn't matter)
    var cfg = require( 'configya' )( { rabbit: { broker: { port: 5672 } } }, './path/to/configuration.json' );

    var port = cfg.rabbit.broker.port; // etc.

The original version of configya used a get method to retrieve configuration values with the ability to specify a default/fallback if the key were missing. This is technically still supported, but we think the new approach (nested keys) is nicer. Here's an example of the original API:

	var config = require( 'configya' )( './path/config.json' );

	// get the value from the config file, if an
	// environment variable is present the environment
	// variable ALWAYS trumps the file setting unless
	// you have deploy-type=DEV in your env settings
	config.get( 'key' );

	config.get( 'key', defaultValue );