|
| 1 | +/* |
| 2 | + * Copyright 2021 Google LLC |
| 3 | + * |
| 4 | + * Licensed under the Apache License, Version 2.0 (the "License"); |
| 5 | + * you may not use this file except in compliance with the License. |
| 6 | + * You may obtain a copy of the License at |
| 7 | + * |
| 8 | + * http://www.apache.org/licenses/LICENSE-2.0 |
| 9 | + * |
| 10 | + * Unless required by applicable law or agreed to in writing, software |
| 11 | + * distributed under the License is distributed on an "AS IS" BASIS, |
| 12 | + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
| 13 | + * See the License for the specific language governing permissions and |
| 14 | + * limitations under the License. |
| 15 | + */ |
| 16 | +package com.google.cloud.pubsublite.flink; |
| 17 | + |
| 18 | +import static com.google.cloud.pubsublite.internal.ExtractStatus.toCanonical; |
| 19 | +import static com.google.cloud.pubsublite.internal.wire.ServiceClients.addDefaultMetadata; |
| 20 | +import static com.google.cloud.pubsublite.internal.wire.ServiceClients.addDefaultSettings; |
| 21 | + |
| 22 | +import com.google.api.core.ApiFutureCallback; |
| 23 | +import com.google.api.core.ApiFutures; |
| 24 | +import com.google.api.gax.rpc.ApiException; |
| 25 | +import com.google.auto.value.AutoValue; |
| 26 | +import com.google.cloud.pubsublite.AdminClient; |
| 27 | +import com.google.cloud.pubsublite.AdminClientSettings; |
| 28 | +import com.google.cloud.pubsublite.Partition; |
| 29 | +import com.google.cloud.pubsublite.SequencedMessage; |
| 30 | +import com.google.cloud.pubsublite.SubscriptionPath; |
| 31 | +import com.google.cloud.pubsublite.cloudpubsub.FlowControlSettings; |
| 32 | +import com.google.cloud.pubsublite.flink.reader.*; |
| 33 | +import com.google.cloud.pubsublite.flink.split.SubscriptionPartitionSplit; |
| 34 | +import com.google.cloud.pubsublite.internal.BlockingPullSubscriber; |
| 35 | +import com.google.cloud.pubsublite.internal.BlockingPullSubscriberImpl; |
| 36 | +import com.google.cloud.pubsublite.internal.CursorClient; |
| 37 | +import com.google.cloud.pubsublite.internal.CursorClientSettings; |
| 38 | +import com.google.cloud.pubsublite.internal.wire.PubsubContext; |
| 39 | +import com.google.cloud.pubsublite.internal.wire.RoutingMetadata; |
| 40 | +import com.google.cloud.pubsublite.internal.wire.SubscriberBuilder; |
| 41 | +import com.google.cloud.pubsublite.internal.wire.SubscriberFactory; |
| 42 | +import com.google.cloud.pubsublite.proto.Cursor; |
| 43 | +import com.google.cloud.pubsublite.proto.SeekRequest; |
| 44 | +import com.google.cloud.pubsublite.v1.SubscriberServiceClient; |
| 45 | +import com.google.cloud.pubsublite.v1.SubscriberServiceSettings; |
| 46 | +import com.google.common.util.concurrent.MoreExecutors; |
| 47 | +import java.io.Serializable; |
| 48 | +import java.util.function.Consumer; |
| 49 | +import java.util.function.Supplier; |
| 50 | +import javax.annotation.Nullable; |
| 51 | +import org.apache.flink.api.connector.source.Boundedness; |
| 52 | +import org.apache.flink.connector.base.source.reader.splitreader.SplitReader; |
| 53 | +import org.apache.flink.util.function.SerializableSupplier; |
| 54 | +import org.slf4j.Logger; |
| 55 | +import org.slf4j.LoggerFactory; |
| 56 | + |
| 57 | +@AutoValue |
| 58 | +public abstract class PubsubLiteSourceSettings<OutputT> implements Serializable { |
| 59 | + private static final Logger LOG = LoggerFactory.getLogger(PubsubLiteSourceSettings.class); |
| 60 | + private static final long serialVersionUID = 3206181560865850636L; |
| 61 | + |
| 62 | + public static <OutputT> Builder<OutputT> builder( |
| 63 | + PubsubLiteDeserializationSchema<OutputT> schema) { |
| 64 | + return new AutoValue_PubsubLiteSourceSettings.Builder<OutputT>() |
| 65 | + .setDeserializationSchema(schema) |
| 66 | + .setBoundedness(Boundedness.CONTINUOUS_UNBOUNDED) |
| 67 | + .setTimestampSelector(MessageTimestampExtractor.publishTimeExtractor()) |
| 68 | + .setPartitionFinishedCondition(PartitionFinishedCondition.continueIndefinitely()); |
| 69 | + } |
| 70 | + |
| 71 | + public static Builder<SequencedMessage> messagesBuilder() { |
| 72 | + return builder(PubsubLiteDeserializationSchema.sequencedMessageSchema()); |
| 73 | + } |
| 74 | + |
| 75 | + // Required |
| 76 | + public abstract SubscriptionPath subscriptionPath(); |
| 77 | + |
| 78 | + // Required |
| 79 | + public abstract FlowControlSettings flowControlSettings(); |
| 80 | + |
| 81 | + // Optional |
| 82 | + public abstract Boundedness boundedness(); |
| 83 | + |
| 84 | + // Optional |
| 85 | + public abstract MessageTimestampExtractor timestampSelector(); |
| 86 | + |
| 87 | + // Optional |
| 88 | + public abstract PartitionFinishedCondition.Factory partitionFinishedCondition(); |
| 89 | + |
| 90 | + // Internal |
| 91 | + abstract PubsubLiteDeserializationSchema<OutputT> deserializationSchema(); |
| 92 | + |
| 93 | + abstract @Nullable SerializableSupplier<AdminClient> adminClientSupplier(); |
| 94 | + |
| 95 | + abstract @Nullable SerializableSupplier<CursorClient> cursorClientSupplier(); |
| 96 | + |
| 97 | + private static SubscriberServiceClient newSubscriberServiceClient( |
| 98 | + SubscriptionPath path, Partition partition) throws ApiException { |
| 99 | + try { |
| 100 | + SubscriberServiceSettings.Builder settingsBuilder = SubscriberServiceSettings.newBuilder(); |
| 101 | + settingsBuilder = |
| 102 | + addDefaultMetadata( |
| 103 | + PubsubContext.of(PubsubContext.Framework.of("FLINK")), |
| 104 | + RoutingMetadata.of(path, partition), |
| 105 | + settingsBuilder); |
| 106 | + return SubscriberServiceClient.create( |
| 107 | + addDefaultSettings(path.location().region(), settingsBuilder)); |
| 108 | + } catch (Throwable t) { |
| 109 | + throw toCanonical(t).underlying; |
| 110 | + } |
| 111 | + } |
| 112 | + |
| 113 | + private static SubscriberFactory getSubscriberFactory( |
| 114 | + SubscriptionPath path, Partition partition, SeekRequest seek) { |
| 115 | + return (consumer) -> |
| 116 | + SubscriberBuilder.newBuilder() |
| 117 | + .setSubscriptionPath(path) |
| 118 | + .setPartition(partition) |
| 119 | + .setServiceClient(newSubscriberServiceClient(path, partition)) |
| 120 | + .setMessageConsumer(consumer) |
| 121 | + .setInitialLocation(seek) |
| 122 | + .build(); |
| 123 | + } |
| 124 | + |
| 125 | + AdminClient getAdminClient() { |
| 126 | + if (adminClientSupplier() != null) { |
| 127 | + return adminClientSupplier().get(); |
| 128 | + } |
| 129 | + return AdminClient.create( |
| 130 | + AdminClientSettings.newBuilder().setRegion(subscriptionPath().location().region()).build()); |
| 131 | + } |
| 132 | + |
| 133 | + CursorClient getCursorClient() { |
| 134 | + if (cursorClientSupplier() != null) { |
| 135 | + return cursorClientSupplier().get(); |
| 136 | + } |
| 137 | + return CursorClient.create( |
| 138 | + CursorClientSettings.newBuilder() |
| 139 | + .setRegion(subscriptionPath().location().region()) |
| 140 | + .build()); |
| 141 | + } |
| 142 | + |
| 143 | + CompletablePullSubscriber.Factory getSplitStateFactory() { |
| 144 | + return split -> { |
| 145 | + SeekRequest seek = |
| 146 | + SeekRequest.newBuilder() |
| 147 | + .setCursor(Cursor.newBuilder().setOffset(split.start().value()).build()) |
| 148 | + .build(); |
| 149 | + SubscriberFactory factory = |
| 150 | + getSubscriberFactory(split.subscriptionPath(), split.partition(), seek); |
| 151 | + |
| 152 | + BlockingPullSubscriber b = new BlockingPullSubscriberImpl(factory, flowControlSettings()); |
| 153 | + return new CompletablePullSubscriberImpl( |
| 154 | + b, partitionFinishedCondition().New(split.subscriptionPath(), split.partition())); |
| 155 | + }; |
| 156 | + } |
| 157 | + |
| 158 | + Supplier<SplitReader<Record<OutputT>, SubscriptionPartitionSplit>> getSplitReaderSupplier() { |
| 159 | + return () -> |
| 160 | + new DeserializingSplitReader<>( |
| 161 | + new MessageSplitReader(getSplitStateFactory()), |
| 162 | + deserializationSchema(), |
| 163 | + timestampSelector()); |
| 164 | + } |
| 165 | + |
| 166 | + Consumer<SubscriptionPartitionSplit> getCursorCommitter() { |
| 167 | + CursorClient client = getCursorClient(); |
| 168 | + return (SubscriptionPartitionSplit split) -> { |
| 169 | + ApiFutures.addCallback( |
| 170 | + client.commitCursor(split.subscriptionPath(), split.partition(), split.start()), |
| 171 | + new ApiFutureCallback<Void>() { |
| 172 | + @Override |
| 173 | + public void onFailure(Throwable throwable) { |
| 174 | + LOG.error("Failed to commit cursor to Pub/Sub Lite ", throwable); |
| 175 | + } |
| 176 | + |
| 177 | + @Override |
| 178 | + public void onSuccess(Void unused) {} |
| 179 | + }, |
| 180 | + MoreExecutors.directExecutor()); |
| 181 | + }; |
| 182 | + } |
| 183 | + |
| 184 | + @AutoValue.Builder |
| 185 | + abstract static class Builder<OutputT> { |
| 186 | + // Required |
| 187 | + public abstract Builder<OutputT> setSubscriptionPath(SubscriptionPath path); |
| 188 | + |
| 189 | + // Required |
| 190 | + public abstract Builder<OutputT> setFlowControlSettings(FlowControlSettings settings); |
| 191 | + |
| 192 | + // Optional |
| 193 | + public abstract Builder<OutputT> setBoundedness(Boundedness value); |
| 194 | + |
| 195 | + // Optional |
| 196 | + public abstract Builder<OutputT> setTimestampSelector(MessageTimestampExtractor value); |
| 197 | + |
| 198 | + // Optional |
| 199 | + public abstract Builder<OutputT> setPartitionFinishedCondition( |
| 200 | + PartitionFinishedCondition.Factory value); |
| 201 | + |
| 202 | + abstract Builder<OutputT> setDeserializationSchema( |
| 203 | + PubsubLiteDeserializationSchema<OutputT> schema); |
| 204 | + |
| 205 | + abstract Builder<OutputT> setAdminClientSupplier(SerializableSupplier<AdminClient> value); |
| 206 | + |
| 207 | + abstract Builder<OutputT> setCursorClientSupplier(SerializableSupplier<CursorClient> value); |
| 208 | + |
| 209 | + abstract PubsubLiteSourceSettings<OutputT> build(); |
| 210 | + } |
| 211 | +} |
0 commit comments