Butter Provider

Butter Project is a toolkit to build VOD platforms, this component is the base class for Providers.

A Provider in the Butter terminology is an accessor for media content, it provides items of type 'movie' or 'tvshow' that can be displayed in a Butter app.

Butter will automatically load any npm package installed (listed in package.json) that matches the /butter-provider-.*/ regex.

Documentation

A Butter Provider is just a npm package that needs to export a specific API that we describe hereafter.

Note that if you want to use the autoload features in Butter you should name your module butter-provider-${something}.

Writing a Provider

We provide a base provider in butter-provider that we recommend extending, it handles caching and a few other bootstrapping quirks, but formally speaking it's not required.

Here we'll be creating a provider for the vodo.net service.

Create a npm module

Create a directory and init a new npm module:

mkdir butter-provider-vodo
cd butter-provider-vodo
yarn init

Depend on butter-provider

First install the npm module and add it as a dependency.

yarn add butter-provider

Then edit your index.js.

'use strict';

const Provider = require('butter-provider');
const defaultConfig = {
    name: 'Vodo',
    tabName: 'Vodo CC Content',
    args: {
        apiUrl: Provider.ArgType.STRING
    },
    defaults: {
        apiUrl: 'http://butter.vodo.net/popcorn'
    },
}

class Vodo extends Provider {
    constructor(args, config = defaultConfig) {
        super(args, config)

        this.apiUrl = this.args.apiUrl
    }
}

Implement required Methods

You need to supply code for at least fetch, other methods like detail, extractIds and resolveStream have default implementations that should work in most cases. That said you probably want to implement those too to be smarter than us.

See the API documentation hereafter for more details.

Use our generic Tests

We have tests for you, to get them running you need to do 2 things;

First add a devDependency on tape and debug so that you can run the tests:

yarn add -D butter-provider-tests

Then you need to tell npm what to run on npm test, and tell the tests how to call your provider using the butter.testArgs key:

  "scripts": {
    "test": "tape ./node_modules/butter-provider/tests/*"
  },
  "butter": {
    "timeout": 20000,
    "testArgs": "vodo?urlList=[ \
        \"http://butter.vodo.net/popcorn\",  \
        \"https://butter.vodo.net/popcorn\", \
        \"http://localhost:8080/popcorn\"    \
    ]"
  },

API

config (Object)

The config object should have the following fields:

Provider.config = {
     name: String,      // will be used in logs to refer to your provider
     tabName: String,   // will appear as the description of the tab
     filters: [Object], // (optional) a list of the filters supported,
                        // see the documentation below
     argTypes: Object || String, // (optional) the args this provider supports
     defaults: Object,  // (optional) default values for the args object
     /* DEPRECATED: should be removed */
     subtitle: String,  // (optional) name of the subtitle provider
     metadata: String   // (optional) name of the metadata provider
     uniqueId: String,  // the name of the field we should use to unify assets
}

filters

The filters object is a collection of mappings between filter-keys and their pretty-printed version in the app. They are used to limit the content in the list view and are passed back to fetch() when modified.

the filters object is of the shape:

const filters = {
    [filterType]: {key: name}
}

We currently support 3 filterTypes:

• genres
• sorters
• types (optional)

Provider.DefaultFilters has definitions for genres and sorters that will be used if none are provided.

argTypes

The argTypes object is a mapping between arg names and Provider.ArgType types, currently these are the supported values:

    Provider.ArgType.ARRAY,   // expects a string of values separated by ',' (JSON.parse)
    Provider.ArgType.OBJECT,  // expects a string of json (JSON.parse)
    Provider.ArgType.BOOLEAN, // expects a string that evaluates to a boolean
    Provider.ArgType.NUMBER,  // expects a string that can be passed to Number()
    Provider.ArgType.STRING,  // expects a string

These types will be automatically instanciated into the args property of the created class, in the long run, there will be UI in Butter to modify those declared args from the settings panels.

It is not required that you use this mechanism (i.e. you can parse your args as you please in your butter-provider) but it will sure save you some headaches.

Note that it can be a string in the uri form, we will then split it with querystring.parse and convert it according to the right Provider.ArgType, this is the mechanism that is in use when you pass this kind of queries:

    vodo?urlList=[ \
        "http://butter.vodo.net/popcorn",  \
        "https://butter.vodo.net/popcorn", \
        "http://localhost:8080/popcorn"    \
    ]

defaults

You can provide default values for the argTypes object. Any JavaScript Object will do.

Note that while we currently do not do type checking on those we may very well start doing so in the future, so make sure the default you provide maps the Provider.ArgType for that element.

fetch (Object: filters -> (promise) Object)

The fetch method is the first called of your provider, it's used to show the content when users open Butter. Keep it small, keep it simple, keep it fast, as load time will depend on performance of fetch. Grab the bare minimum of data you need, you'll have other opportunities to enrich that data in subsequent calls (like detail or resolveStream).

The fetch method takes in a set of filters that can have the following keys:

const filters = {
    keywords: [String],          // keywords to search for
    genre: String,               // limit to this genre
    order: Provider.OrderType,   // sort order (ASC, DESC)
    sorter: Provider.SorterType, // sorter type (NAME, RATING, POPULARITY)
    limit: Number,               // number of elements to return
}

with Provider.OrderType being:

    Provider.Ordertype.ASC,    // items are sorted in ascending order
    Provider.Ordertype.DESC,   // items are sorted in descending order
    Provider.OrderType.NULL    // items are not sorted

with Provider.SorterType being:

    Provider.Sortertype.NAME,        // items are sorted by name
    Provider.Sortertype.RATING,      // items are sorted by rating
    Provider.Sortertype.POPULARITY,  // items are sorted by popularity
    Provider.SorterType.NULL         // items are not sorted

The fetch method returns a promise that resolves to an object of the shape:

const fetchReturn = {
    results: [Object],  // returned result items
    hasMore: Boolean    // can the provider get more results ?
}

The results items can have any shape but are required to have at least:

const result = {
    id: String,                // the unique id
    title: String,             // title of the asset
    year: Number,              // year of release
    genres: [String],          // a free list of genre keywords
    rating: Number,            // a 0-100 rating value
    poster: String,            // url to the poster image
    type: Provider.ItemType,   // used by the browser to decide how to show the item
                               // (MOVIE, TVSHOW)
    num_seasons: Number        // the number of seasons available to display
                               // only in the case of Provider.ItemType.TVSHOW
}

Provider.ItemType can be one of:

    Provider.ItemType.MOVIE,  // item is a Movie
    Provider.ItemType.TVSHOW  // item is a TvShow

detail (String: id, Object old_data -> (promise) Object)

The detail method allows you to fetch more metadata from your API when presenting a specific asset, it returns a result object as described in fetch, and takes the id and the data returned by previous fetch calls as an argument.

It is important, to split detail and fetch data gets, as it can be heavy on your API endpoint to get all those details at once.

Note that the expected shape of detail results are slightly different for Provider.ItemType.MOVIE and Provider.ItemType.TVSHOW.

Note that the default implementation will just return the object untouched, you don't need to implement a function like:

detail(torrent_id, old_data) {
    return Promise.resolve(old_data);
};

The required info is:

const detail = {
    id: String,                // the unique id
    title: String,             // title of the asset
    year: Number,              // year of release
    genres: [String],          // a free list of genre keywords
    rating: Number,            // a 0-100 rating value
    poster: String,            // url to the poster image
    type: Provider.ItemType    // used by the browser to decide how to show the item
                               // (MOVIE, TVSHOW)
    num_seasons: Number,       // the number of seasons available to show
                               // only in the case of Provider.ItemType.TVSHOW
//--- all of the above is already required by fetch(), new stuff here: ---
    runtime: Number,           // the runtime of the item
    backdrop: String,          // url to the backdrop image
    subtitle: {url: String},   // language -> url subs mapping
    synopsis: String,          // a short description of the asset
}

Provider.ItemType.MOVIE

The Provider.ItemType.MOVIE has the following additional fields:

const detail = {
//--- including all the fields of the generic detail object
    sources: Object,          // sources object
    trailer: String           // url of the trailer, formatted for butter-streamers
}

Provder.ItemType.TVSHOW

The Provider.ItemType.TVSHOW also has a few additional fields to include:

const detail = {
//--- including all the fields of the generic detail object
    status: String,              // the status of the item
    episodes: [Object],          // the episodes to display
}

The episodes array will have the following shape:

{
    sources: Object               // a sources Object
    watched: Boolean              // indication if an episode has been watched
    first_aired: Number,          // epoch time when the episode was first aired
    overview: String,             // small description of the episode
    episode: Number,              // episode number of the season
    season: String,               // season number of the episode
    tvdb_id: Number               // the tvdb id of the episode
}

sources Object

The end goal of these methods is to return sources objects that have the following shape:

Note that the 'torrent' name is a bit confusing and really a legacy name, it should be called 'resources'.

const sources = {
    [Provider.QualityType]: {     // the quality of the episode
        url: String,              // the resource's url, formatted for butter-streamers
        size: Number,             // the resource's descriptor size (magnet/torrent/hls playlist)
                                  // only for Provider.ItemType.MOVIE
        filesize: String          // (optional) the resource's main video filesize
        peers: Number,            // (optional) number of seeds
        seeds: Number,            // (optional) number of peers
    }
}

With Provider.QualityType being:

    Provider.QualityType.DEFAULT   // The default object to stream
    Provider.QualityType.LOW       // 480p quality
    Provider.QualityType.MEDIUM    // 720p quality
    Provider.QualityType.HIGH      // 1080p quality

extractIds ([Object]: items -> [String])

This method is used to keep a cache of the content in a Butter app. The generic implementation is:

extractIds(items) {
    return _.pluck(items.results, 'id');
};

(optional) resolveStream (src, config, data -> (promise) String)

This method is used to let the provider decide what the end url should be according to some config passed by the apps. It's main purpose is to allow the selection of different languages, but in the future it may allow for deeper customizations (as for instance choosing a streaming technology).

The default handler will just return src that is the legacy value providers are required to return in fetch and details for torrent data.

Currently config will have this shape:

{
  audio: String,
}

data will be whatever data was returned from the latest fetch or details for the current media, it is given raw so that you can control where to 'hide' the urls you will want to switch on languages switches.

(optional) random (void -> (promise) Object)

Returns a random result item as described in detail.

(optional) update (void -> (promise) [Object])

Allows to notify the Provider it can update it's internal cache (not used).