Navigator Prefabricating Marinates

    nuc

    0.5.2 • Public • Published

    NUC

    Build Status Dependency Status npm version Coverage Status

    Configuration via JSON and INI file hierarchy: cli, env, local, user, system, global.

    Uses specified app/module id and process.platform value to look for config files. Also uses the CLI args and environment values.

    Provides a flexible way to specify configuration values in multiple system places in an intuitive hierarchy. This allows users and systems to configure values for apps/modules in simple text files as well as override them with CLI args.

    Also provides a nuc command, installed globally, to review configuration values and set them at specific levels in the hierarchy.

    Install

    # Install global for `nuc` command. 
    npm install -g nuc
     
    # Install locally to get configuration for your module/app 
    npm install --save nuc

    Usage: the values object

    The primary use of nuc is to read a hierarchy of configuration files and combine them along the way into a single object. The final object contains the "configuration" information we want.

    I call that the "values" object. It's produced by the "collapse" mode which is enabled by default.

    // require module, call it, give it the ID and some defaults
    // id is required, defaults are optional.
    // values contains all key/value pairs combined into one object.
    // later object key/value pairs override earlier ones
    var nuc = require('nuc')
      , values = nuc({ id:'appid', defaults:{} })

    Usage: the ValueStore stack

    See ValueStore for a description of that module.

    The nuc module can gather all the configuration files into a ValueStore instance. Each configuration source remains separate and its hierarchy is maintained.

    Each one knows its "source". If it's a cli option then its source is 'cli'. Environment values are from 'env'. Files have a file property with the path.

    The various sources in the ValueStore can be inspected in many ways. Check out ValueStore for more on that.

    To receive only the ValueStore you must enable it and disable the "collapse mode".

    var nuc = require('nuc')
      , store = nuc({ id:'someid', collapse:false, stack:true  });

    Usage: Have it both ways

    By default, nuc only produces the single values object and returns it. I call that "collapse mode". By default, the ValueStore isn't produced because the "stack mode" is off.

    To have both, enable both, and the returned result will contain both a values property and a store property.

    Because "collapse" is enabled by default, just enable "stack" and you get both.

    var nuc = require('nuc')
      , both = nuc({ id:'someid', defaults:{}, collapse:true, stack:true });

    Usage: CLI

    The nuc module has a cli tool also named "nuc". You may use it locally or install it globally.

    It has multiple subcommands which you may recognize from other configuration tools such as npm config for "npmrc".

    One major difference between nuc and all the others is nuc requires an "id" be specified which is used in many ways, one of them is as part of the file names. This means different applications may have their own configuration file inside a user's home directory. They are separate.

    For example, a user may have a configuration folder in their home directory at /home/username/conf. Inside it, there may be many files which will be loaded by nuc when their corresponding app "id" is used.

    The help information and usage displayed by the command itself is more descriptive than the documentation here. Install it and check it out :)

    The first three commands in the list below are informational tasks which don't require an app "id". The later ones all require an app "id" to perform. The app "id" can be specified on the command line each time as the first arg to nuc, and, it must have an @ symbol in front of it.

    For example:

    nuc @someid list each
    nuc @someid get somekey
    nuc @someid add somekey somevalue

    It may become tedious typing the "id" for each command. As a shortcut I've added the ability to specify the "id" in a file. When you run nuc commands it will look in the current working directory for a file named '.nuc.name'. If it finds it, it will read its contents expecting to find a single word which it will use as the app "id".

    For example, if the PWD had a .nuc.name file like:

    theapp

    Then, when you're in that same directory, you can leave out the app "id" from the command and it will use that. For example, now you can do:

    nuc list each
    nuc get somekey
    nuc add somekey somevalue

    It will use "theapp" as the app "id" for those commands the same as if you typed "@theapp" as the second arg to the nuc command.

    The cli subcommands are:

    1. version - prints the version to the console
    2. usage - prints the commands which can be used with nuc. This is the default subcommand. (alias = 'use')
    3. help - prints general help info, or, info about specific things when passed as an arg, such as nuc help list or nuc help scope
    4. list - list the contents of the values object or of the store object. You may limit it to a single scope instead of all scopes. Uses util.inspect to add colors to the output. (alias = 'show').
    5. get - get the value for a key for the first scope where the key exists, or, for the scope you specify.
    6. set - sets the key/value pair you specify into the first configuration file found within all writable scopes, or the one scope you specify. If no file is found, then a default file location is used to write out a new file with the key/value pair. It defaults to a path in the user's home directory where they're sure to have permission to write a file. These values may be overridden by using nuc itself to set their configuration values with app id "nuc". Because, nuc uses itself for its own configuration.
    7. add - similar to set except it combines your new value and the existing value into an array. If the key doesn't exist yet, then this command is the same as set.
    8. remove - the opposite of set, it will remove the key from the first scope it is found. You may specific the scope to remove it from, or, scope 'all' to remove it from all scopes. (As I write this, I realize I haven't completed the "remove from all" functionality yet.)

    TODO: list the values used by nuc by default and what the keys are to override them.

    These are not all the commands. Just a minimum of examples...

    nuc version
     
    nuc usage
     
    # there's more help info than these... 
    nuc help
    nuc help usage
    nuc help list
    nuc help scope
     
    # list all scopes together, or, a single scope, or each scope separately 
    nuc @someid list
    nuc @someid list local
    nuc @someid list each
     
    # get a key's value from anywhere it's found, or look in a specific scope... 
    nuc @someid get somekey
    nuc @someid get somekey user
     
    # set into an existing file, or create a new one, or into a specific scope 
    nuc @someid set somekey value
    nuc @someid set key2 value2 local
     
    # key 'somekey' will then have the value: [ 'value', 'value2' ] 
    nuc @someid add somekey value2
     
    # remove a key from the first scope it's found in, or, from a specific scope 
    nuc @someid remove somekey
    nuc @someid remove somekey user
     
    # there's more... 

    Value Hierarchy

    Values are loaded from multiple scopes. Items higher in this list override those below them.

    1. cli - command line arguments override all other sources of values. Provided in process.argv. These are read-only.
    2. env - environment values are prepared by the OS and override all but CLI args. Provided in node as process.env. These are read-only.
    3. local - find config files in the local current working directory and up the path to root. Their accessibility depend on the user permissions.
    4. user - look in user's home directory. These are read/write accessible.
    5. system - system configuration directories, such as /etc and c:\ProgramData. These values are writable if the user has access to them.
    6. global - where modules are installed with npm install -g. These values are read-only.

    Keys

    The id specified to nuc is used as the base name to produce keys and file names.

    1. cli key is the uppercase form of the id, prepended with two dashes, followed by an underscore, then the keys, an equal sign, and the value, like:

      --ID_KEY=value

    Keys can be nested by using two underscores together. For example:

    // key: --ID_ONE__TWO__THREE=value is interpreted as:
    result = {
      ONE: {
        TWO: {
          THREE: 'value'
        }
      }
    }
    1. env keys are the same as cli keys, without the '--' in front: process.env.ID_ONE__TWO__THREE.

    File names

    Many files are searched for with various patterns using the id in it. See below for complete list.

    On Windows platforms:

    1. look for ini files.
    2. env.USERPROFILE is used instead of env.HOME for 'user' scope
    3. env.ALLUSERPROFILE is used instead of /etc for 'system' scope
    4. ??? what should be used instead of /usr/lib and /usr/local/lib for 'global' ?

    On *nix platforms:

    1. look for dot files (filename starts with a dot)
    2. look for files suffixed with rc, like: .idrc
    3. look in /etc, /usr/lib/node_modules, and /usr/local/lib/node_modules

    Using 'id' for the id these are all the files searched for:

    NOTE: This is an outdated list... but, it gives you an idea of it, so, I'll leave it for now. Look at windows-bases, nix-bases, and path-generator to get a better idea.

    1. '/usr/lib/node_modules/id/id.conf'
    2. '/usr/lib/node_modules/id/id.json'
    3. '/usr/lib/node_modules/id/.idrc'
    4. '/usr/lib/node_modules/id/idrc'
    5. '/usr/local/lib/node_modules/id/id.conf'
    6. '/usr/local/lib/node_modules/id/id.json'
    7. '/usr/local/lib/node_modules/id/.idrc'
    8. '/usr/local/lib/node_modules/id/idrc'
    9. '/etc/id.conf'
    10. '/etc/id.json'
    11. '/etc/.idrc'
    12. '/etc/idrc'
    13. '/etc/id/id.conf'
    14. '/etc/id/id.json'
    15. '/etc/id/.idrc'
    16. '/etc/id/idrc'
    17. env.HOME + '/id.ini'
    18. env.HOME + '/id.conf'
    19. env.HOME + '/id.json'
    20. env.HOME + '/.idrc'
    21. env.HOME + '/idrc'
    22. env.HOME + '/id/id.ini'
    23. env.HOME + '/id/id.conf'
    24. env.HOME + '/id/id.json'
    25. env.HOME + '/id/.idrc'
    26. env.HOME + '/id/idrc'
    27. CWD + '/id.ini'
    28. CWD + '/id.conf'
    29. CWD + '/id.json'
    30. CWD + '/.idrc'
    31. CWD + '/idrc'
    32. CWD + '/id/id.ini'
    33. CWD + '/id/id.conf'
    34. CWD + '/id/id.json'
    35. CWD + '/id/.idrc'
    36. CWD + '/id/idrc'

    Why name it nuc ?

    Chance, mostly. I began the nup project and wanted to read my configuration from multiple places on the system. I reviewed a bunch of modules providing configuration methods. The one I liked the most was one I saw long ago when I first moved to node: the npmrc stuff. I decided I wanted to make a module I could use with any app or module I made.

    So, I needed a name for this new project coming directly from the nu (nup) project. Add a 'c' for configuration, and, it sounds like "nook", which, may be where we find all the configuration information, in the "nook and crannies"?

    I basically wanted to stop thinking about a name and begin writing the code. So, nuc it is.

    MIT License

    Install

    npm i nuc

    DownloadsWeekly Downloads

    9

    Version

    0.5.2

    License

    MIT

    Last publish

    Collaborators

    • elidoran