Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

2.x: add Flowable.parallel() and parallel operators #4974

Merged
merged 2 commits into from
Jan 18, 2017
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
95 changes: 95 additions & 0 deletions src/main/java/io/reactivex/Flowable.java
Original file line number Diff line number Diff line change
Expand Up @@ -29,6 +29,7 @@
import io.reactivex.internal.schedulers.ImmediateThinScheduler;
import io.reactivex.internal.subscribers.*;
import io.reactivex.internal.util.*;
import io.reactivex.parallel.ParallelFlowable;
import io.reactivex.plugins.RxJavaPlugins;
import io.reactivex.schedulers.*;
import io.reactivex.subscribers.*;
Expand Down Expand Up @@ -10363,6 +10364,100 @@ public final Flowable<T> onTerminateDetach() {
return RxJavaPlugins.onAssembly(new FlowableDetach<T>(this));
}

/**
* Parallelizes the flow by creating multiple 'rails' (equal to the number of CPUs)
* and dispatches the upstream items to them in a round-robin fashion.
* <p>
* Note that the rails don't execute in parallel on their own and one needs to
* apply {@link ParallelFlowable#runOn(Scheduler)} to specify the Scheduler where
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

What about remove runOn and add Scheduler as a parameter to parallel()?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It is the same logic as with regular factory methods such as just, range, fromIterable don't take a Scheduler, plus you can apply multiple runOn's on a sequence at different stages. For example create a pipeline with stages of parallelism=2 and 3 stages in total.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

plus you can apply multiple runOn's on a sequence at different stages

Ah, that's nice, got it.

* each rail will execute.
* <p>
* To merge the parallel 'rails' back into a single sequence, use {@link ParallelFlowable#sequential()}.
* <p>
* <img width="640" height="310" src="https://raw.github.com/wiki/ReactiveX/RxJava/images/rx-operators/flowable.parallel.png" alt="">
* <dl>
* <dt><b>Backpressure:</b></dt>
* <dd>The operator requires the upstream to honor backpressure and each 'rail' honors backpressure
* as well.</dd>
* <dt><b>Scheduler:</b></dt>
* <dd>{@code parallel} does not operate by default on a particular {@link Scheduler}.</dd>
* </dl>
* @return the new ParallelFlowable instance
* @since 2.0.5 - experimental
*/
@BackpressureSupport(BackpressureKind.FULL)
@SchedulerSupport(SchedulerSupport.NONE)
@CheckReturnValue
@Experimental
public final ParallelFlowable<T> parallel() {
return ParallelFlowable.from(this);
}

/**
* Parallelizes the flow by creating the specified number of 'rails'
* and dispatches the upstream items to them in a round-robin fashion.
* <p>
* Note that the rails don't execute in parallel on their own and one needs to
* apply {@link ParallelFlowable#runOn(Scheduler)} to specify the Scheduler where
* each rail will execute.
* <p>
* To merge the parallel 'rails' back into a single sequence, use {@link ParallelFlowable#sequential()}.
* <p>
* <img width="640" height="310" src="https://raw.github.com/wiki/ReactiveX/RxJava/images/rx-operators/flowable.parallel.png" alt="">
* <dl>
* <dt><b>Backpressure:</b></dt>
* <dd>The operator requires the upstream to honor backpressure and each 'rail' honors backpressure
* as well.</dd>
* <dt><b>Scheduler:</b></dt>
* <dd>{@code parallel} does not operate by default on a particular {@link Scheduler}.</dd>
* </dl>
* @param parallelism the number of 'rails' to use
* @return the new ParallelFlowable instance
* @since 2.0.5 - experimental
*/
@BackpressureSupport(BackpressureKind.FULL)
@SchedulerSupport(SchedulerSupport.NONE)
@CheckReturnValue
@Experimental
public final ParallelFlowable<T> parallel(int parallelism) {
ObjectHelper.verifyPositive(parallelism, "parallelism");
return ParallelFlowable.from(this, parallelism);
}

/**
* Parallelizes the flow by creating the specified number of 'rails'
* and dispatches the upstream items to them in a round-robin fashion and
* uses the defined per-'rail' prefetch amount.
* <p>
* Note that the rails don't execute in parallel on their own and one needs to
* apply {@link ParallelFlowable#runOn(Scheduler)} to specify the Scheduler where
* each rail will execute.
* <p>
* To merge the parallel 'rails' back into a single sequence, use {@link ParallelFlowable#sequential()}.
* <p>
* <img width="640" height="310" src="https://raw.github.com/wiki/ReactiveX/RxJava/images/rx-operators/flowable.parallel.png" alt="">
* <dl>
* <dt><b>Backpressure:</b></dt>
* <dd>The operator requires the upstream to honor backpressure and each 'rail' honors backpressure
* as well.</dd>
* <dt><b>Scheduler:</b></dt>
* <dd>{@code parallel} does not operate by default on a particular {@link Scheduler}.</dd>
* </dl>
* @param parallelism the number of 'rails' to use
* @param prefetch the number of items each 'rail' should prefetch
* @return the new ParallelFlowable instance
* @since 2.0.5 - experimental
*/
@BackpressureSupport(BackpressureKind.FULL)
@SchedulerSupport(SchedulerSupport.NONE)
@CheckReturnValue
@Experimental
public final ParallelFlowable<T> parallel(int parallelism, int prefetch) {
ObjectHelper.verifyPositive(parallelism, "parallelism");
ObjectHelper.verifyPositive(prefetch, "prefetch");
return ParallelFlowable.from(this, parallelism, prefetch);
}

/**
* Returns a {@link ConnectableFlowable}, which is a variety of Publisher that waits until its
* {@link ConnectableFlowable#connect connect} method is called before it begins emitting items to those
Expand Down
Original file line number Diff line number Diff line change
@@ -0,0 +1,165 @@
/**
* Copyright (c) 2016-present, RxJava Contributors.
*
* Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in
* compliance with the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software distributed under the License is
* distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See
* the License for the specific language governing permissions and limitations under the License.
*/

package io.reactivex.internal.operators.parallel;

import java.util.concurrent.Callable;

import org.reactivestreams.*;

import io.reactivex.exceptions.Exceptions;
import io.reactivex.functions.BiConsumer;
import io.reactivex.internal.subscribers.DeferredScalarSubscriber;
import io.reactivex.internal.subscriptions.*;
import io.reactivex.parallel.ParallelFlowable;
import io.reactivex.plugins.RxJavaPlugins;

/**
* Reduce the sequence of values in each 'rail' to a single value.
*
* @param <T> the input value type
* @param <C> the collection type
*/
public final class ParallelCollect<T, C> extends ParallelFlowable<C> {

final ParallelFlowable<? extends T> source;

final Callable<C> initialCollection;

final BiConsumer<C, T> collector;

public ParallelCollect(ParallelFlowable<? extends T> source,
Callable<C> initialCollection, BiConsumer<C, T> collector) {
this.source = source;
this.initialCollection = initialCollection;
this.collector = collector;
}

@Override
public void subscribe(Subscriber<? super C>[] subscribers) {
if (!validate(subscribers)) {
return;
}

int n = subscribers.length;
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

final

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Same in other similar places would be good.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I don't do those unless the variable has to be accessed from an inner class.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

In long methods reader has to spend extra time to check that it's not modified anywhere, but ok

@SuppressWarnings("unchecked")
Subscriber<T>[] parents = new Subscriber[n];

for (int i = 0; i < n; i++) {

C initialValue;

try {
initialValue = initialCollection.call();
} catch (Throwable ex) {
Exceptions.throwIfFatal(ex);
reportError(subscribers, ex);
return;
}

if (initialValue == null) {
reportError(subscribers, new NullPointerException("The initialSupplier returned a null value"));
return;
}

parents[i] = new ParallelCollectSubscriber<T, C>(subscribers[i], initialValue, collector);
}

source.subscribe(parents);
}

void reportError(Subscriber<?>[] subscribers, Throwable ex) {
for (Subscriber<?> s : subscribers) {
EmptySubscription.error(ex, s);
}
}

@Override
public int parallelism() {
return source.parallelism();
}

static final class ParallelCollectSubscriber<T, C> extends DeferredScalarSubscriber<T, C> {


private static final long serialVersionUID = -4767392946044436228L;

final BiConsumer<C, T> collector;

C collection;

boolean done;

ParallelCollectSubscriber(Subscriber<? super C> subscriber,
C initialValue, BiConsumer<C, T> collector) {
super(subscriber);
this.collection = initialValue;
this.collector = collector;
}

@Override
public void onSubscribe(Subscription s) {
if (SubscriptionHelper.validate(this.s, s)) {
this.s = s;

actual.onSubscribe(this);

s.request(Long.MAX_VALUE);
}
}

@Override
public void onNext(T t) {
if (done) {
return;
}

try {
collector.accept(collection, t);
} catch (Throwable ex) {
Exceptions.throwIfFatal(ex);
cancel();
onError(ex);
return;
}
}

@Override
public void onError(Throwable t) {
if (done) {
RxJavaPlugins.onError(t);
return;
}
done = true;
collection = null;
actual.onError(t);
}

@Override
public void onComplete() {
if (done) {
return;
}
done = true;
C c = collection;
collection = null;
complete(c);
}

@Override
public void cancel() {
super.cancel();
s.cancel();
}
}
}
Original file line number Diff line number Diff line change
@@ -0,0 +1,88 @@
/**
* Copyright (c) 2016-present, RxJava Contributors.
*
* Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in
* compliance with the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software distributed under the License is
* distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See
* the License for the specific language governing permissions and limitations under the License.
*/

package io.reactivex.internal.operators.parallel;

import org.reactivestreams.*;

import io.reactivex.functions.Function;
import io.reactivex.internal.functions.ObjectHelper;
import io.reactivex.internal.operators.flowable.FlowableConcatMap;
import io.reactivex.internal.util.ErrorMode;
import io.reactivex.parallel.ParallelFlowable;

/**
* Concatenates the generated Publishers on each rail.
*
* @param <T> the input value type
* @param <R> the output value type
*/
public final class ParallelConcatMap<T, R> extends ParallelFlowable<R> {

final ParallelFlowable<T> source;

final Function<? super T, ? extends Publisher<? extends R>> mapper;

final int prefetch;

final ErrorMode errorMode;

public ParallelConcatMap(
ParallelFlowable<T> source,
Function<? super T, ? extends Publisher<? extends R>> mapper,
int prefetch, ErrorMode errorMode) {
this.source = source;
this.mapper = ObjectHelper.requireNonNull(mapper, "mapper");
this.prefetch = prefetch;
this.errorMode = ObjectHelper.requireNonNull(errorMode, "errorMode");
}

@Override
public int parallelism() {
return source.parallelism();
}

@Override
public void subscribe(Subscriber<? super R>[] subscribers) {
if (!validate(subscribers)) {
return;
}

int n = subscribers.length;

@SuppressWarnings("unchecked")
final Subscriber<T>[] parents = new Subscriber[n];

// FIXME cheat until we have support from RxJava2 internals
Publisher<T> p = new Publisher<T>() {
int i;

@SuppressWarnings("unchecked")
@Override
public void subscribe(Subscriber<? super T> s) {
parents[i++] = (Subscriber<T>)s;
}
};

FlowableConcatMap<T, R> op = new FlowableConcatMap<T, R>(p, mapper, prefetch, errorMode);

for (int i = 0; i < n; i++) {

op.subscribe(subscribers[i]);
// FIXME needs a FlatMap subscriber
// parents[i] = FlowableConcatMap.createSubscriber(s, mapper, prefetch, errorMode);
}

source.subscribe(parents);
}
}
Loading