== Connect Configuration Loader

This provides:

• a method to externalise configuration from one or more outside configuration YAML files,
• supports layering, so each file can get more specific (start with common, override with environmental)
• allows that configuration to be transformed by registered initializers (eg. Vault keywords being referenced which get resolved automatically once the config is loaded and inserted in place of the keys).
• loads the environment into the configuration under the prefix "env", so you can approach all of your configuration in a consistent manner
• works from the command line or via an environment variable (CONNECT_CONFIG)
• allows you to arbitrarily get config out with a path and a default
• allows you to inject config from a tree node into an object, including subobjects
• gives you first class support for annotating configuration in Typescript using 'configKey'
• gives you a postConfigured function that can be called on your object when configuration has been injected.
• watches for config file changes and repushes their contents to the objects registered with set (aka Kubernetes)

Note: this library is asynchronous because initializers can be asynchronous (e.g. Vault).

==== Defining Configuration Files

If an environment variable called CONNECT_CONFIG exists, it will use this, otherwise it will look on the command line. In both cases, it looks for -p file.yml -p file2.yml ... - a chain of -p followed by a space followed by a filename. It will load those in in order, with values from the second, etc files replacing the same named fields in the first.

If there are no configuration files, it does not fire.

To load the configuration you must call "load" on the configurationLoader that is exported from the library.

[source,javascript]

const configLoader = require('connect-config-loader').configurationLoader; await configLoader.load();

==== Environment variables

Once all of the config has been copied in, it will copy the system environment variables into the configuration under the prefix env. So env.CONNECT_CONFIG will exist if you used the environment variable for configuration as per above.

==== Manually getting data

This you can use to get values if you wish:

[source,javascript]

configLoader.get('env.CONNECT_CONFIG', 'some-default-value');

You can also set a variable and ask it to inject all elements under that tree into your object:

[source,javascript]

const env = {_configIndex: 'env', postConfigured: () => { console.log('finished!'); }}; configLoader.set(env); // should print "finished!" if you change your config while this is running it will do so each time

==== Typescript

With Typescript you can use annotations instead, and give the option of the field being required and having alternative keys:

class MyConfigClass { @configKey('env.DB_PASSWORD', true, ['db.password']) public String dbPassword;

public postConfigured() { logger.info('have db info'); } }

const mcc = new MyConfigClass(); configLoader.set(mcc);

It is possible to do this for Javascript by pushing the config into the Object, but its quite painful.