npm:minima | Skypack

README

Minima

The blog engine where you can do what you want

What?

With the surge of text editors and cloud storage why can't I use any of them to write my blog? Why do I have to use some crappy HTML5 editor when I have native apps that give me rich features? Why do I have to use your crappy native app to write my blog? Why can't I use plain text markdown? Why can't I use WYSIWYG? How can I trust your document storage to keep my blogs? Maybe I'll keep them in Dropbox, Google Drive, Evernote, etc as a backup.

These are the questions I've heard and asked over and over again. Almost every time I find a cool new blog engine I go through this. I always feel tied to whatever editor and storage the blog engine decides on.

minima |ˈminəmə|
plural form of minimum.

Minima aims to be a middle man and a host. The purpose is to let you write however you please. You can write your blogs in anything you want (Evernote, Dropbox, Google Docs). This also opens the door to any text editor you wish. Once you decide how you want to write, find a loader and tell Minima to use it. Can't find one for your middleware? Then write one!

Choosing a Loader

Once you decide on an article loader, mark it as an npm dependency and tell minima to use it. Do this by modifying the app.config.js file.

{
  //...

  articleMiddleware: 'minima-evernote-loader',

  //...

  'minima-evernote-loader': {
      notebookName: 'minima',
      key: '...',
      secret: '...',
      sandbox: false
  },

  //...
}

How to Write Middleware

Examples:

How you write your middleware is irrelevant. The only requirements is that you exports an Object (which will be constructed) and expose certain methods for Minima to use. They are as follows:

connect
connectCallback
listPages
getPageContent

Constructor

Your constructor will be passed three things: logger, config, and store.

Signature: function(logger, config, store) {}

logger - an instance of log4js. It is highly recommended that you use this so that debugging issues doesn't become an issue in the future. All logging methods use util.format because readability.
config - object containing the configuration values for your middleware. Also contains callbackUrl for oauth purposes (which will directly invoke the connectCallback method of your middleware.
store - a key/value store which exposes get(key), set(key, value), and remove(key). It is sandboxed from the rest of the application. Use this sparingly (i.e. storing oauth tokens).

connect

Signature: function(req, res, next){}

Called from the '/admin/connect' page. When the user first loads the application they will be directed to connect their middleware. Middleware.connect is called directly from Express. You have full access to the request and response objects. This might change in the future. In this method you should perform your typical oauth functions, passing the config.callbackUrl so that the oauth provider invokes your middleware's connectCallback function.

connectCallback

Signature: function(req, res, next) {}

This is only necessary if you're authorization process requires a callback URL (i.e oauth). In your connect method you can use the config.callbackUrl to give to your oauth provider. Once the oauth passes the provider should invoke this method. Once you've completed the oauth you should persist the authentication into the store and attempt to retrieve it in your constructor.

listPages

Signature: function(callback) {}

callback: a Callback function to be invoked with each page. When invoking the callback you should pass one single object which should match the following:

{
    identifier: // some unique identifier that your middleware can identify this note by,
    title: // the title to be shown on the page
    slug: // the slug to use in the url for this page (will default to [slug][] the title)
    createDate: // Date (defaults to Date.now)
    modifiedDate: // Date (defaults to Date.now) and used to determine if an update is needed
    tags: // tags for the page
    unpublished: // boolean - passed if the article should be unpublished
}

getPageContent

Only called when a page detects an update. This prevents Minima from having to load every page each time the application starts. Sometimes it is not possible for your middleware to work this way (e.g. Dropbox). If this is the case then you should cache the pages content when you receive it during the listPages method and fetch that cached content during getPageContent.

Signature: function(note) {}

note - an object containing everything received from listPages

Returns:

Q.promise

Promise should be resolved with the FULL NOTE. Not just the content. This is necessary because the receiving end of this callback is stateless and does not retain each call it makes to getPageContent. When the promise is resolved it will depend on the Note object in it's entirety.

To see implementations of middleware see the examples listed above.

Usage no npm install needed!