Kotlin Spotify Web API
This is the Kotlin implementation of the Spotify Web API
Table of Contents
- Library installing
- Documentation
- Need help, have a question, or want to contribute?
- Creating a new api instance
- Tips
- Notes
- Contributing
Library installing
JVM, Android, JS
repositories {
jcenter()
}
compile group: 'com.adamratzman', name: 'spotify-api-kotlin-core', version: '3.2.14'
Note that images and profiles are not supported on the Kotlin/JS target.
Android
If you declare any release types not named debug or release, you may see "Could not resolve com.adamratzman:spotify-api-kotlin-android:VERSION". You need to do the following for each release type not named debug or release:
android {
buildTypes {
yourReleaseType1 {
// ...
matchingFallbacks = ['release', 'debug']
}
yourReleaseType2 {
// ...
matchingFallbacks = ['release', 'debug']
}
...
}
}
To successfully build, you might need to exclude kotlin_module files from the packaging. To do this, inside the android/buildTypes/release closure, you would put:
packagingOptions {
exclude 'META-INF/*.kotlin_module'
}
Documentation
The spotify-web-api-kotlin
JavaDocs are hosted at https://adamint.github.io/spotify-web-api-kotlin/spotify-web-api-kotlin/
Have a question?
If you have a question, you can:
- Create an issue
- Join our Discord server
- Contact me using Adam#9261 on Discord
Creating a new api instance
To decide which api you need (SpotifyAppApi, SpotifyClientApi, SpotifyImplicitGrantApi), you can refer to the sections below or the Spotify authorization guide. In general:
- If you don't need client resources, use SpotifyAppApi
- If you're using the api in a backend application, use SpotifyClientApi (with or without PKCE)
- If you're using the api in Kotlin/JS browser, use SpotifyImplicitGrantApi
- If you need access to client resources in an Android or other application, use SpotifyClientApi with PKCE
SpotifyAppApi
This provides access only to public Spotify endpoints. Use this when you have a server-side application. Note that implicit grant authorization provides a higher api ratelimit, so consider using implicit grant if your application has significant usage.
By default, the SpotifyApi Token
automatically regenerates when needed. This can be changed by overriding the automaticRefresh
builder setting.
There are four exposed builders, depending on the level of control you need over api creation. Please see the spotifyAppApi builder docs for a full list of available builders.
You will need:
- Spotify application client id
- Spotify application client secret
Example creation (default settings)
val api = spotifyAppApi("clientId", "clientSecret").build() // create and build api
println(api.browse.getNewReleases().complete()) // use it
Example creation, using an existing Token and setting automatic token refresh to false
val token = spotifyAppApi(spotifyClientId, spotifyClientSecret).build().token
val api = spotifyAppApi(
"clientId",
"clientSecret",
token,
SpotifyApiOptionsBuilder(
automaticRefresh = false
)
)
println(api.browse.getNewReleases().complete()) // use it
SpotifyClientApi
The SpotifyClientApi
is a superset of SpotifyApi
; thus, nothing changes if you want to access public data. This library does not provide a method to retrieve the code from your callback url; instead, you must implement that with a web server. Automatic refresh is available only when building with an authorization code or a Token
object. Otherwise, it will expire Token.expiresIn
seconds after creation.
Make sure your application has requested the proper Scopes in order to ensure proper function of this library. The api option requiredScopes
allows you to verify that a client has actually authorized with the scopes you are expecting.
You will need:
- Spotify application client id
- Spotify application client secret (if not using PKCE)
- Spotify application redirect uri
- To choose which client authorization method (PKCE or non-PKCE) to use
PKCE
Use the PKCE builders and helper methods if you are using the Spotify client authorization PKCE flow. Building via PKCE returns a SpotifyClientApi
which has modified refresh logic.
Use cases:
- You are using this library in an application (likely Android), or do not want to expose the client secret.
To learn more about the PKCE flow, please read the Spotify authorization guide. Some highlights about the flow are:
- It is refreshable, but each refresh token can only be used once. This library handles token refresh automatically by default
- It does not require a client secret; instead, a set redirect uri and a random code verifier are used to verify the authenticity of the authorization.
- A code verifier is required. The code verifier is "a cryptographically random string between 43 and 128 characters in length. It can contain letters, digits, underscores, periods, hyphens, or tildes."
- A code challenge is required. "In order to generate the code challenge, your app should hash the code verifier using the SHA256 algorithm. Then, base64url encode the hash that you generated."
- When creating a pkce api instance, the code verifier is passed in by you and compared to the code challenge used to authorize the user.
This library contains helpful methods that can be used to simplify the PKCE authorization process. This includes getSpotifyPkceCodeChallenge
(not available in the Kotlin/JS target), which SHA256 hashes and base64url encodes the code challenge, and getPkceAuthorizationUrl
, which allows you to generate an easy authorization url for PKCE flow.
Please see the spotifyClientPkceApi builder docs for a full list of available builders.
Takeaway: Use PKCE authorization flow in applications where you cannot secure the client secret.
To get a PKCE authorization url, to which you can redirect a user, you can use the getPkceAuthorizationUrl
top-level method. An example is shown below, requesting 4 different scopes.
val codeVerifier = "thisisaveryrandomalphanumericcodeverifierandisgreaterthan43characters"
val codeChallenge = getSpotifyPkceCodeChallenge(codeVerifier) // helper method
val url: String = getPkceAuthorizationUrl(
SpotifyScope.PLAYLIST_READ_PRIVATE,
SpotifyScope.PLAYLIST_MODIFY_PRIVATE,
SpotifyScope.USER_FOLLOW_READ,
SpotifyScope.USER_LIBRARY_MODIFY,
clientId = "clientId",
redirectUri = "your-redirect-uri",
codeChallenge = codeChallenge
)
There is also an optional parameter state
, which helps you verify the authorization.
Note: If you want automatic token refresh, you need to pass in your application client id and redirect uri when using the spotifyClientPkceApi
.
Example: A user has authorized your application. You now have the authorization code obtained after the user was redirected back to your application. You want to create a new SpotifyClientApi
.
val codeVerifier = "thisisaveryrandomalphanumericcodeverifierandisgreaterthan43characters"
val code: String = ...
val api = spotifyClientPkceApi(
"clientId", // optional. include for token refresh
"your-redirect-uri", // optional. include for token refresh
code,
codeVerifier, // the same code verifier you used to generate the code challenge
SpotifyApiOptionsBuilder(
retryWhenRateLimited = false
)
).build()
println(api.library.getSavedTracks().complete().take(10).filterNotNull().map { it.track.name })
Non-PKCE (backend applications, requires client secret)
To get a non-PKCE authorization url, to which you can redirect a user, you can use the getSpotifyAuthorizationUrl
top-level method. An example is shown below, requesting 4 different scopes.
val url: String = getSpotifyAuthorizationUrl(
SpotifyScope.PLAYLIST_READ_PRIVATE,
SpotifyScope.PLAYLIST_MODIFY_PRIVATE,
SpotifyScope.USER_FOLLOW_READ,
SpotifyScope.USER_LIBRARY_MODIFY,
clientId = "clientId",
redirectUri = "your-redirect-uri",
state = "your-special-state" // optional
)
There are also several optional parameters, allowing you to set whether the authorization url is meant for implicit grant flow, the state, and whether a re-authorization dialog should be shown to users.
There are several exposed builders, depending on the level of control you need over api creation. Please see the spotifyClientApi builder docs for a full list of available builders.
Example: You've redirected the user back to your web server and have an authorization code (code).
In this example, automatic token refresh is turned on by default.
val authCode = ""
val api = spotifyClientApi(
"clientId",
"clientSecret",
"your-redirect-uri",
authCode
).build() // create and build api
println(api.personalization.getTopTracks(limit = 5).complete().items.map { it.name }) // print user top tracks
Example: You've saved a user's token from previous authorization and need to create an api instance.
In this case, if you provide a client id to the builder, automatic token refresh will also be turned on.
val token: Token = ... // your existing token
val api = spotifyClientApi(
"clientId",
"clientSecret",
"your-redirect-uri",
token,
SpotifyApiOptionsBuilder(
onTokenRefresh = {
println("Token refreshed at ${System.currentTimeMillis()}")
}
)
).build()
println(api.personalization.getTopTracks(limit = 5).complete().items.map { it.name })
SpotifyImplicitGrantApi
Use the SpotifyImplicitGrantApi
if you are using the Spotify implicit grant flow. SpotifyImplicitGrantApi
is a superset of SpotifyClientApi
. Unlike the other builders, the spotifyImplicitGrantApi
builder method directly returns a SpotifyImplicitGrantApi
instead of an api builder.
Use cases:
- You are using the Kotlin/JS target for this library.
- Your frontend Javascript passes the token received through the implicit grant flow to your backend, where it is then used to create an api instance.
To learn more about the implicit grant flow, please read the Spotify authorization guide. Some highlights about the flow are:
- It is non-refreshable
- It is client-side
- It does not require a client secret
Please see the spotifyImplicitGrantApi builder docs for a full list of available builders.
The Kotlin/JS target contains the parseSpotifyCallbackHashToToken
method, which will parse the hash for the current url into a Token object, with which you can then instantiate the api.
Takeaway: There are two ways to use implicit grant flow, browser-side only and browser and server. This library provides easy access for both.
Example
val token: Token = ...
val api = spotifyImplicitGrantApi(
null,
null,
token
) // create api. there is no need to build it
println(api.personalization.getTopArtists(limit = 1).complete()[0].name) // use it
SpotifyApiBuilder Block & setting API options
There are three pluggable blocks in each api's corresponding builder
credentials
lets you set the client id, client secret, and redirect uriauthorization
lets you set the type of api authorization you are using. Acceptable types include: an authorization code, aToken
object, a Token's access code string, and an optional refresh token stringoptions
lets you configure API options to your own specific needs
API options
This library does not attempt to be prescriptivist. All API options are located in SpotifyApiOptions
and their default values can be overridden; however, use caution in doing so, as most of the default values either allow for significant performance or feature enhancements to the API instance.
useCache
: Set whether to cache requests. Default: truecacheLimit
: The maximum amount of cached requests allowed at one time. Null means no limit. Default: 200automaticRefresh
: Enable or disable automatic refresh of the Spotify access token when it expires. Default: trueretryWhenRateLimited
: Set whether to block the current thread and wait until the API can retry the request. Default: trueenableLogger
: Set whether to enable to the exception logger. Default: truetestTokenValidity
: After API creation, test whether the token is valid by performing a lightweight request. Default: falsedefaultLimit
: The default amount of objects to retrieve in one request. Default: 50json
: The Json serializer/deserializer instance.allowBulkRequests
: Allow splitting too-large requests into smaller, allowable api requests. Default: truerequestTimeoutMillis
: The maximum time, in milliseconds, before terminating an http request. Default: 100000msrefreshTokenProducer
: Provide if you want to use your own logic when refreshing a Spotify token.requiredScopes
: Scopes that your application requires to function (only applicable toSpotifyClientApi
andSpotifyImplicitGrantApi
). This verifies that the token your user authorized with actually contains the scopes your application needs to function.
Notes:
- Unless you have a good reason otherwise,
useCache
should be true cacheLimit
is per Endpoint, not per API. Don't be surprised if you end up with over 200 items in your cache with the default settings.automaticRefresh
is disabled when client secret is not provided, or if tokenString is provided in SpotifyClientApiallowBulkRequests
for example, lets you query 80 artists in one wrapper call by splitting it into 50 artists + 30 artistsrefreshTokenProducer
is useful when you want to re-authorize with the Spotify Auth SDK or elsewhere
Using the API
APIs available in all SpotifyApi
instances, including SpotifyClientApi
and SpotifyImplicitGrantApi
:
SearchApi
(searching items)AlbumApi
(get information about albums)BrowseApi
(browse new releases, featured playlists, categories, and recommendations)ArtistApi
(get information about artists)PlaylistApi
(get information about playlists)UserApi
(get public information about users on Spotify)TrackApi
(get information about tracks)FollowingApi
(check whether users follow playlists)
APIs available only in SpotifyClientApi
and SpotifyImplicitGrantApi
instances:
ClientSearchApi
(all the methods inSearchApi
, and searching shows and episodes)EpisodeApi
(get information about episodes)ShowApi
(get information about shows)ClientPlaylistApi
(all the methods inPlaylistApi
, and get and manage user playlists)ClientProfileApi
(all the methods inUserApi
, and get the user profile, depending on scopes)ClientFollowingApi
(all the methods inFollowingApi
, and get and manage following of playlists, artists, and users)ClientPersonalizationApi
(get user top tracks and artists)ClientLibraryApi
(get and manage saved tracks and albums)ClientPlayerApi
(view and control Spotify playback)
Tips
Building the API
The easiest way to build the API is synchronously using .build() after a builder
spotifyAppApi(clientId, clientSecret).build()
You can also build the API asynchronously using Kotlin coroutines.
runBlocking {
spotifyAppApi(clientId, clientSecret).buildAsyncAt(this) { api ->
// do things
}
}
What is the SpotifyRestAction class?
Abstracting requests into a SpotifyRestAction
class allows for a lot of flexibility in sending and receiving requests. This class includes options for asynchronous and blocking execution in all endpoints. However, due to this, you must call one of the provided methods in order for the call to execute! The SpotifyRestAction
provides many methods and fields for use, including blocking and asynchronous ones. For example,
hasRun()
tells you whether the rest action has been startedhasCompleted()
tells you whether this rest action has been fully executed and completedcomplete()
blocks the current thread and returns the resultsuspendComplete(context: CoroutineContext = Dispatchers.Default)
switches to given context, invokes the supplier, and synchronously retrieves the result.suspendQueue()
suspends the coroutine, invokes the supplier asynchronously, and resumes with the resultqueue()
executes and immediately returnsqueue(consumer: (T) -> Unit)
executes the provided callback as soon as the request is asynchronously evaluatedqueueAfter(quantity: Int, timeUnit: TimeUnit, consumer: (T) -> Unit)
executes the provided callback after the provided time. As long as supplier execution is less than the provided time, this will likely be accurate within a few milliseconds.asFuture()
transforms the supplier into aCompletableFuture
(only JVM)
SpotifyRestPagingAction
Separate from, but a superset of SpotifyRestAction, this specialized implementation of RestActions includes extensions for AbstractPagingObject
(PagingObject
and CursorBasedPagingObject
). This class gives you the same functionality as SpotifyRestAction, but you also have the ability to retrieve all of its items or linked PagingObjects, or a subset of its items or linked PagingObjects with one call, with a single method call to getAllItems()
or getAllPagingObjects()
, or getWithNext(total: Int, context: CoroutineContext = Dispatchers.Default)
or getWithNextItems(total: Int, context: CoroutineContext = Dispatchers.Default)
respectively
Notes
The benefits of LinkedResults, PagingObjects, and Cursor-based Paging Objects
Spotify provides these three object models in order to simplify our lives as developers. So let's see what we can do with them!
PagingObjects
PagingObjects are a container for the requested objects (items
), but also include important information useful in future calls. It contains the request's limit
and offset
, along with (sometimes) a link to the next and last page of items and the total number of items returned.
If a link to the next or previous page is provided, we can use the getNext
and getPrevious
methods to retrieve the respective PagingObjects
Cursor-Based Paging Objects
A cursor-based paging object is a PagingObject with a cursor added on that can be used as a key to find the next page of items. The value in the cursor, after
, describes after what object to begin the query.
Just like with PagingObjects, you can get the next page of items with getNext
. However, there is no provided implementation of after
in this library. You will need to do it yourself, if necessary.
LinkedResults
Some endpoints, like PlaylistAPI.getPlaylistTracks
, return a LinkedResult, which is a simple wrapper around the list of objects. With this, we have access to its Spotify API url (with href
), and we provide simple methods to parse that url.
Generic Request
For obvious reasons, in most cases, making asynchronous requests via queue
or queueAfter
is preferred. However, the synchronous format is also shown.
val api = spotifyAppApi(
System.getenv("SPOTIFY_CLIENT_ID"),
System.getenv("SPOTIFY_CLIENT_SECRET")
).build()
// block and print out the names of the twenty most similar songs to the search
println(api.search.searchTrack("Début de la Suite").complete().joinToString { it.name })
// now, let's do it asynchronously
api.search.searchTrack("Début de la Suite").queue { tracks -> println(tracks.joinToString { track -> track.name }) }
// simple, right? what about if we want to print out the featured playlists message from the "Overview" tab?
println(api.browse.getFeaturedPlaylists().complete().message)
// easy! let's try something a little harder
// let's find out Bénabar's Spotify ID, find his top tracks, and print them out
val benabarId = api.search.searchArtist("Bénabar").complete()[0].id
// this works; a redundant way would be: api.artists.getArtist("spotify:artist:6xoAWsIOZxJVPpo7Qvqaqv").complete().id
println(api.artists.getArtistTopTracks(benabarId).complete().joinToString { it.name })
Track Relinking
Spotify keeps many instances of most tracks on their servers, available in different markets. As such, if we use endpoints that return tracks, we do not know if these tracks are playable in our market. That's where track relinking comes in.
To relink in a specified market, we must supply a market
parameter for endpoints where available. In both Track and SimpleTrack objects in an endpoint response, there is a nullable field called linked_from
. If the track is unable to be played in the specified market and there is an alternative that is playable, this will be populated with the href, uri, and, most importantly, the id of the track.
You can then use this track in SpotifyClientApi
endpoints such as playing or saving the track, knowing that it will be playable in your market!
Contributing
See CONTRIBUTING.md