Rxjava 介紹與 Android 中的 RxJava

RxJava 介紹與
Android 中的 RxJava
⿈黃千碩 (Kros)
oSolve Ltd. / Wish8 Co,. Ltd.
Mobile App Developer

Outline
• RxJava Introduction
• Observable
• Operators
• Subject
• Scheduler
• Android Lifecycle
• Testing
• Performance

What is RxJava?
• RxJava 是 Reactive X (Reactive Extensions) 對
Java VM 的實作
• 是⼀一個 Library，利⽤用資料流 (observable
sequences) 來處理「asynchronous 與 event-
base 」類型的程式。
• 主要⾓角⾊色為：observable 與 observer 。

What is Reactive?
• 翻譯：「響應式」「反應式」開發
• 把資料 (data) 或是事件 (event) 變成「可觀察」
(observer pattern) 的「資料流」。
• 並加上運算元 (operators) 來操作這些資料。

What is FRP ?
• FRP - Functional Reactive Programming.
• Reactive 是⺫⽬目的
• 為了能讓開發者不落⼊入如何處理（事件）資料的
繁雜程式邏輯中，利⽤用函數式 (Functional) 的⽅方
法來處理資料流
• ﬁlter(), map(), ﬂatMap(), …etc.

Why FRP?
• Concurrency
• thread 的控管複雜
• Asynchronous Programming
• 為追求 60fps，許多事情我們會丟到背景處理
• Callback Hell
• 當 Callback 太多時，眼睛都花了。

– 林信良
“若開發者是以務實且不斷提升作為⾃自我期許，更
⾼高階的抽象化作法將會是必修的課題”

Observable
• Observable 為發射資料的⼈人

Create Observable
• Observable.just()
• Observable.from()
• …etc.

Observable.just()
• 把「資料」轉變成 Observable。
Observable.just("Hello World!")

Observer
• Observer 為對這些資料有興趣的⼈人
• 透過 subscribe method 連結 observer 與 observable.
• Observer 透過 subscribe 來監聽⼀一個 Observable.

Subscribe
• 連結 observable 與 observer
• 通常必須實作 subscribe 的 interface.
• onNext, onError, onComplete
public final Subscription subscribe(final Action1<? super T> onNext,
final Action1<Throwable> onError,
final Action0 onComplete) {
/* ... */
}

>>>>>>>>>>>>>>>>>>> s:Hello World!
Observable.just("Hello World!").subscribe(new Action1<String>() { 
@Override 
public void call(String s) { 
System.out.println(">>>>>>>>>>>>>>>>>>> s:" + s); 
} 
}, new Action1<Throwable>() { 
@Override 
public void call(Throwable throwable) { 
} 
}, new Action0() { 
@Override 
public void call() { 
} 
});

>>>>>>>>>>>>>>>>>>> s:Hello World!
@Override 
} 
@Override 
} 
@Override 
} 
});
@Override 
} 
});
可以只實作感興趣的 callback

>>>>>>>>>>>>>>>>>>> s:Hello World!
@Override 
} 
@Override 
} 
@Override 
} 
});
@Override 
} 
});
Observable.just("Hello World!").subscribe(s -> { 
});
套⽤用 retrolambda，採⽤用 java8 lambda，讓程式碼更簡潔

Observable.from
• 把「⼀一包資料」轉變成 Observable。⽽而這個
Observable 每次只發射資料中的單⼀一資料
Observable.from(listOfIntegers)

Observable.from
>>>>>>>>>>>>>>>>>>> integer:1
>>>>>>>>>>>>>>>>>>> integer:2
>>>>>>>>>>>>>>>>>>> integer:3
>>>>>>>>>>>>>>>>>>> integer:4
>>>>>>>>>>>>>>>>>>> integer:5
>>>>>>>>>>>>>>>>>>> integer:6
>>>>>>>>>>>>>>>>>>> integer:7
List<Integer> integers = new ArrayList<>(); 
integers.add(1); 
// ... 
integers.add(7);
 
Observable.from(integers).subscribe(integer -> { 
System.out.println(">>>>>>>>>>>>>>>>>>> integer:" + integer); 
});

“Hot” and “Cold”
Observable
• Observable 什麼時候會發射資料呢？
• Hot observable
• 當它⼀一建⽴立時就會發射資料
• Cold observable
• 當有 observer subscribe 時，才會發射資料

Operators
• Creating Observables (ex: create, from, just, …)
• Transforming Observables (ex: map, ﬂatMap, …)
• Filtering Observables
• Combining Observables
• Error Handling Operators
• Observable Utility Operators
• ……etc.

Observable.just("Hello World!").map(s -> s + " Android Taipei") 
.subscribe(s -> { 
});
Observable.from(integers) 
.map(integer -> integer + 10) 
.subscribe(integer -> { 
System.out.println(">>>>>>>>>>>>>>>>>>> integer:" + integer); 
});
>>>>>>>>>>>>>>>>>>> s:Hello World! Android Taipei
>>>>>>>>>>>>>>>>>>> integer:11
>>>>>>>>>>>>>>>>>>> integer:12
>>>>>>>>>>>>>>>>>>> integer:13
>>>>>>>>>>>>>>>>>>> integer:14
>>>>>>>>>>>>>>>>>>> integer:15
>>>>>>>>>>>>>>>>>>> integer:16
>>>>>>>>>>>>>>>>>>> integer:17
對 "Hello World!" 加⼯工
對 list 中每個 element 加⼯工

台北市
公廁查詢系統
利⽤用台北市政府 Open Data
http://data.taipei/
⺫⽬目的：找尋離我最近的公廁

// - 跟 Server 抓取公廁資料
@GET("/apiAccess") 
void listToiletCallback(@Query("rid") String rid,  
@Query("scope") String scope, 
@Query("limit") int limit, 
@Query("offset") int offset, 
Callback<ApiResponse> callback);

@Override 
public void onViewCreated(View view, Bundle savedInstanceState) { 
super.onViewCreated(view, savedInstanceState); 
progressBar.setVisibility(View.VISIBLE);
fetchNearestToilet();
}

@Override 
progressBar.setVisibility(View.VISIBLE);
fetchNearestToilet();
}
⺫⽬目的：找出距離我 5 km 以內的公廁，
並按照遠近排序

private void fetchNearestToilet() { 
apiService.listToiletCallback(RID, SCOPE, 500, 0, new Callback<ApiResponse>() { 
@Override 
public void success(ApiResponse apiResponse, Response response) { 
List<Toilet> filtered = new ArrayList<>(); 
for (Toilet toilet : apiResponse.getResult().getToilets()) { 
if (lessThan5Km(toilet)) { 
filtered.add(toilet); 
} 
} 
Collections.sort(filtered, new Comparator<Toilet>() { 
@Override 
public int compare(Toilet lhs, Toilet rhs) { 
return compareDistance(lhs, rhs); 
} 
}); 
adapter.reset(filtered); 
progressBar.setVisibility(View.GONE); 
} 
 
@Override 
public void failure(RetrofitError error) { 
progressBar.setVisibility(View.GONE); 
ViewHelper.showError(getActivity(), error); 
} 
}); 
}

Observable<ApiResponse> listToilet(@Query("rid") String rid,  
@Query("scope") String scope,  
@Query("limit") int limit,  
@Query("offset") int offset);

@Override 
progressBar.setVisibility(View.VISIBLE); 
fetchNearestToilet() 
.observeOn(AndroidSchedulers.mainThread()) 
.finallyDo(() -> progressBar.setVisibility(View.GONE)) 
.subscribe((toilets) -> { 
adapter.reset(toilets); 
}, 
throwable -> ViewHelper.showError(getActivity(), throwable));
}
}

@Override 
.subscribe(adapter::reset, 
}
Java 8 的 method reference

@Override 
}

private Observable<List<Toilet>> fetchNearestToilet() { 
return apiService.listToilet(RID, SCOPE, 500, 0) 
.flatMap(response -> Observable.from(response.getResult().getToilets())) 
.filter(this::lessThan5Km) 
.toSortedList(this::compareDistance); 
}
@Override 
throwable -> ViewHelper.showError(getActivity(), throwable)); 
}
}

Subject
• 翻譯：主題
• A Subject is a sort of bridge or proxy that is
available in some implementations of
ReactiveX that acts both as an observer and
as an Observable.
• Subject 可以是發送 event 的⼈人 (observable)，
也可以是註冊 event 的⼈人 (observer)。
• ⽤用途：Event Bus

Subject
• Subject 有很多種：
• AsyncSubject
• BehaviorSubject
• PublishSubject
• ReplaySubject

Publish Subject
• 會發送給每個 observers
• 只會接收到 subscribe 之後的 event

Activity 1
Subject
subject.subscribe();

Activity 1
Subject
startActivity();

Activity 1 Activity 2
Subject
startActivity();

Subject
startActivity();
// Do something…
subject.onNext(Event);
finish();

Subject
startActivity();
// Do something…
finish();

Activity 2
Subject
startActivity();
// Do something…
finish();
Activity 1
(with Event)

Activity 1
(with Event)
Activity 2
Subject
startActivity();
// Do something…
finish();

Subject
Activity 1
(with Event)

Subject
Activity 1
(with Event)
Activity 2
(with Event)
Activity 3
(with Event)
可以很多⼈人註冊

Scheduler
• If you want to introduce multithreading into
your cascade of Observable operators, you
can do so by instructing those operators (or
particular Observables) to operate on
particular Schedulers.
• 可以利⽤用 Scheduler 來實作 thread 的切換。

Scheduler
@Override 
}
}

Android Lifecycle
• Activity 與 Fragment 都有各⾃自的 lifecycle.
• Activity, onCreate(), onResume(), onPause(),
onDestory(), ..etc.
• Fragment, onCreate(), onCreateView(), onResume(),
onPause(), onDestory(), ..etc
• 如果 Activity/Fragment 被 destroy 時，你的 async task
還沒做完怎麼辦？

Android Lifecycle
• 會導致 Memory leak 或是 NPE.
• Activity 與 Fragment 都有各⾃自的 lifecycle.
• Activity, onCreate(), onResume(), onPause(),
onDestory(), ..etc.
• Fragment, onCreate(), onCreateView(), onResume(),
onPause(), onDestory(), ..etc
• 如果 Activity/Fragment 被 destroy 時，你的 async task
還沒做完怎麼辦？

Android Lifecycle
• Import RxJava Android 版 
compile 'io.reactivex:rxandroid:0.25.0'
• 使⽤用 Android 相關的 observable 與 event. 
rx.android.lifecycle.LifecycleObservable 
rx.android.lifecycle.LifecycleEvent

@Override 
protected void onStart() { 
super.onStart(); 
lifecycleSubject.onNext(LifecycleEvent.START); 
} 
 
@Override 
protected void onResume() { 
super.onResume(); 
lifecycleSubject.onNext(LifecycleEvent.RESUME); 
} 
 
@Override 
protected void onPause() { 
lifecycleSubject.onNext(LifecycleEvent.PAUSE); 
super.onPause(); 
} 
 
@Override 
protected void onStop() { 
lifecycleSubject.onNext(LifecycleEvent.STOP); 
super.onStop(); 
} 
 
@Override 
protected void onDestroy() { 
lifecycleSubject.onNext(LifecycleEvent.DESTROY); 
super.onDestroy(); 
}
private final BehaviorSubject<LifecycleEvent> lifecycleSubject = BehaviorSubject.create();
public class BaseActivity extends AppCompatActivity {
}

public class BaseActivity extends AppCompatActivity {
/* reset code */
public Observable<LifecycleEvent> lifecycle() { 
return lifecycleSubject.asObservable(); 
}
protected <T> Observable<T> bind(Observable<T> observable) { 
return LifecycleObservable.bindActivityLifecycle(lifecycle(), 
observable.observeOn(AndroidSchedulers.mainThread())); 
}
/* reset code */
}

@Override 
bind(fetchNearestToilet()) 
}
}
@Override 
}
}
bind() 的功能：當 fragment 被 destroyed 時，會⾃自動
unsubscribe 此 observable.

@Override 
}
}
@Override 
bind(fetchNearestToilet()) 
}
}
bind() 的功能：當 fragment 被 destroyed 時，會⾃自動
unsubscribe 此 observable.

Testing
• 測試容易
• 簡易的⽅方法, toBlocking()
• 正規的⽅方法, TestSubscriber()

public class AccountDaemon {
public Observable<Account> login(final Account account) { 
return Observable.just(account).map(account1 -> { 
checkAccount(account); 
return accountService.login(account); 
}); 
}
private void checkAccount(Account account) throws IllegalArgumentException { 
if (TextUtils.isEmpty(account.getEmail()) 
|| TextUtils.isEmpty(account.getPassword())) { 
throw new IllegalArgumentException("Email or password can not be
empty."); 
} 
}
}
public class Account { 
private final String email; 
private final String password; 
public static Account createLoginAccount(final String email,  
final String password) { 
return new Account(email, password); 
} 
// rest implementation…
}

public void testLogin_empty_email() throws Exception { 
Account account = Account.createLoginAccount(null, "password"); 
try { 
accountDaemon.login(account).toBlocking().single(); 
fail("method should throw exception"); 
} catch (Throwable ex) { 
assertEquals("Email or password can not be empty.", ex.getLocalizedMessage()); 
} 
}
// Official way
public void testLogin_using_test_subscriber() { 
TestSubscriber<Account> testSubscriber = new TestSubscriber<>(); 
Account account = Account.createLoginAccount("email", "password"); 
accountDaemon.login(account).subscribe(testSubscriber); 
 
Account expect = Account.createLoginAccount("email", "password"); 
testSubscriber.assertNoErrors(); 
testSubscriber.assertValue(expect); 
}
// Blocking way
public void testLogin() throws Exception { 
Account account = Account.createLoginAccount("email", "password"); 
Account result = accountDaemon.login(account).toBlocking().single(); 
assertEquals("email", result.getEmail()); 
assertEquals("password", result.getPassword()); 
}

public void testPerformance_rx() { 
List<Integer> data = new ArrayList<>(); 
for (int i = 0; i < 100000; ++i) { 
data.add(i); 
} 
List<Integer> result = Observable.from(data) 
.filter(integer -> integer % 2 == 0).toList().toBlocking().first(); 
assertEquals(100000 / 2, result.size()); 
}
public void testPerformance_for_loop() { 
List<Integer> data = new ArrayList<>(); 
for (int i = 0; i < 100000; ++i) { 
data.add(i); 
} 
List<Integer> result = new ArrayList<>(); 
for (int i = 0, size = data.size(); i < size; i++) { 
if (i % 2 == 0) { 
result.add(i); 
} 
} 
assertEquals(100000 / 2, result.size()); 
}

不是每個語⾔言的 Rx 版本效
能都很好，導⼊入時需注意
例如:ReactiveCocoa

RAC-ReactiveCocoa
- (void)testRACPerformance { 
NSArray *array = [self getTestArray]; 
[self measureBlock:^{ 
RACSequence *sequence = [array.rac_sequence filter:^BOOL(NSNumber *number) { 
return number.intValue % 2 == 0; 
}]; 
NSArray *results = sequence.array; 
XCTAssertEqualObjects(@(100000 / 2), @(results.count)); 
}]; 
}
- (void)testNativePerformance { 
NSArray *array = [self getTestArray]; 
[self measureBlock:^{ 
NSMutableArray *results = [NSMutableArray array]; 
for (int i = 0; i < array.count; ++i) { 
NSNumber *number = array[i]; 
if (number.intValue % 2 == 0) { 
[results addObject:number]; 
} 
} 
XCTAssertEqualObjects(@(100000 / 2), @(results.count)); 
}]; 
}

優缺點
• 優點
• 程式碼清楚，簡潔
• 容易進⾏行 Asynchronous Programming
• 缺點
• 學習成本⾼高(map????, ﬂatMap?????, amb???)
• ⼊入侵式的，所有 API 被迫改成 Observable<T>

public void fetchUserProfile() { 
// code 
} 
 
public void fetchFriends() { 
// code 
} 
 
public void fetchShippingInfo() { 
// code 
}
public Observable<Profile> fetchUserProfile() { 
// code 
} 
 
public Observable<List<Friend>> fetchFriends() { 
// code 
} 
 
public Observable<ShippingInfo> fetchShippingInfo() { 
// code 
}

Reference
• ReactiveX 
http://reactivex.io/
• FRP與函數式-林信良 
http://www.ithome.com.tw/voice/91328
• RxJava Android Patterns 
http://stablekernel.com/blog/replace-asynctask-asynctaskloader-rx-
observable-rxjava-android-patterns/
• Architecting Android…The evolution 
http://fernandocejas.com/2015/07/18/architecting-android-the-evolution/
• Unit Testing RxJava Observables 
https://medium.com/ribot-labs/unit-testing-rxjava-6e9540d4a329
• Demo Project 
https://github.com/ch8908/rxjava-demo

Rxjava 介紹與 Android 中的 RxJava

More Related Content

What's hot

Similar to Rxjava 介紹與 Android 中的 RxJava

Recently uploaded

In this document

Rxjava 介紹與 Android 中的 RxJava