Layered config for giants with axes
npm install flapjacksflapjacks is a configuration management module. Its primary goal is to make keeping separate configurations for different environments easy and convenient while making configuration itself as flexible as possible.
* Stackable configuration based on environment name, machine name, user name, and main module name
* Helpers for deep access into the configuration object hierarchy without null check boilerplate
* Opt-in global configuration to any nested node_modules use the first one loaded
* Support for JS, JSON, and relaxed JSON configs.
``js`
var config = require('flapjacks').read();
var someVar = config.get('some.nested.path', []); // if that path is undefined, the [] is returned
if (config.foo) {
// the properties can also be accessed directly, if that's your thing
}
flapjacks looks for files based on process.NODE_ENV, current user, host name, and main module name. It looks in ., ./config/, ~/.$module/, and ~/.config/$module/, where $module is the main module. So if the main module is MyNodeThing, the paths would be ~/.config/MyNodeThing/ and ~/.MyNodeThing/. You can specify the main module name with the environment variable FLAPJACK_NAME. You can also specify a name in the options parameter of the read function, otherwise, it will be the name from the first package.json found from the working directory up.
The files are loaded in order from least to most specific:
*
*
*
*
*
*
*
*
*
with extensions applied as such:
* .json
*
*
*
*
*
*
This paths searched are in order from least to most specific as well:
* ./
*
*
*
*
Each matching config file that is found will be loaded in order, so general settings may be set in ./config/$module.default.config.js and user specific settings could be set in ~/.config/$module/$user.config.js.
To prevent accidental config overlap, files for paths that aren't specific to the main module will have the module name prepended. Where the default config for $module in its ~/.config/$module/ directory would be ~/.config/$module/default.config, it would be ~/.$module.default.config in the home directory and ~/.config/$module.default.config in the ~/.config directory.
Note that the least specific file name is an empty string, which means that extensions like config, rjson, etc end up being the file name. This is handy though, when the module name is automatically prepended e.g. ~/.$module.rjson.
The contents of each config file are loaded as the body to a function:
`js`
function(config) {
// your config code here
}
The function is executed and passed a config object, which is also the context. The same config object will be passed through each config file, so more specific settings will override less specific settings.
Any Node.js functionality is available within the config function, so you can do anything that a regular node script can do within the function. Be careful with how you run a program using this! To be fair though, if you have a compromised account or codebase, this is probably the least of your worries.
Relaxed JSON config files are simply JavaScript files that are expected to contain a top-level object definition. They may contain comments and any other valid JS syntax with the only requirement being that there be a single top level set of matching { and } and no comments before the opening { that also contain a {. The contents of an RJSON file are also used as the body of a function, but the first { will be replaced with return { first.
This format is handy for letting you skip all of the usual JSON boilerplate while also letting you specify functions in your config object.
The result of the RJSON config will be merged into the root of the config.
JSON files are parsed with JSON.parse and the result is merged into the root of the config.
options should be an object with any of the following keys:name
* - Read config files for the module named name or the main module name for this process if name was not specified. The configuration will be cached, so subsequent calls to read will yield the cached result.skipRoot[root]
* - where root is one of CWD, CWDConfig, Config, HomeModule, or Home - don't look in the named path for config filesskipAllRoots
* - don't look in any of the default paths for config filesroots
* - an array of additional root paths to search for config filesskipType[type]
* - where type is one of JSON, RJSON, or Script - don't look for config files of the given type. This is handy to set JSON as the only type of file parsed in case you don't want any eval'ed config code.
Throws away any cached config and reads the config files again with the given options.
Parses the config from content instead of pulling it from config files. This is mostly useful for testing. The type parameter is optional and may be json, rjson, or js. The default type is js.
If bool is true, then a global variable is set with the current instance of flapjacks so that any other modules that get loaded in the process will use this instance instead of getting their own. If it is false, the global variable is emptied and everybody gets their own box of toys again.
After the config has been read, the cached version is accessible as the module property config.
The config object that is passed around and subsequently returned to the caller has a few helper methods that make setting nested variables and handling arrays a little less painful.
Note: path is always a string with . separated keys.
js
set(path, value);
set(settings);
`When called with a path and value, the value at path will become
. If there are missing intermediate levels, they will be created for you.When called with a settings object, each key/value pair will be handled as if it was passed to
set(key, value).When called with an empty path
'', the value will be merged into the root config.$3
`js
merge(path, value);
`Merge the object
into the given path. This is useful for adding settings to an existing setting hierarchy.$3
`js
get(path[, default]);
` traverses to the given path and returns any value found there. This can be used to return intermediate objects branches as well as value leaves. If an intermiediate does not exist, the optional default (otherwise undefined) will be returned instead.When called with an empty path
'', the config object is returned.$3
`js
has(path);
` traverses to the given path and returns true if there is a value (even if that value is specifically set to undefined). If there is no value at any point in the path, has will return false.$3
`js
ensure(path, value);
` traverses to the given path, and if it, or any intermediate levels, do not have a value set, then the given value will be set for the path.$3
`js
push(path, value...);
pop(path[, default]);
splice(path, index, drop, add...);
shift(path[, default]);
unshift(path, value...);
`Each of these does what you'd expect from the corresponding
version after looking up the array at path. If the value at path is truthy but not an array, an exception will be thrown. If the value at path is falsey, a new array will be added for push, unshift, and splice and the optional default (otherwise undefined) will be returned from shift and pop`.