lukecwik commented on a change in pull request #16130: URL: https://github.com/apache/beam/pull/16130#discussion_r766161342
########## File path: sdks/java/harness/src/main/java/org/apache/beam/fn/harness/Caches.java ########## @@ -0,0 +1,285 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one + * or more contributor license agreements. See the NOTICE file + * distributed with this work for additional information + * regarding copyright ownership. The ASF licenses this file + * to you 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 org.apache.beam.fn.harness; + +import java.util.Arrays; +import java.util.Collections; +import java.util.Objects; +import java.util.Set; +import java.util.function.Function; +import org.apache.beam.sdk.options.PipelineOptions; +import org.apache.beam.sdk.options.SdkHarnessOptions; +import org.apache.beam.vendor.guava.v26_0_jre.com.google.common.annotations.VisibleForTesting; +import org.apache.beam.vendor.guava.v26_0_jre.com.google.common.collect.Sets; +import org.apache.beam.vendor.guava.v26_0_jre.com.google.common.util.concurrent.MoreExecutors; +import org.cache2k.Cache2kBuilder; +import org.cache2k.operation.Weigher; +import org.github.jamm.MemoryMeter; + +/** Utility methods used to instantiate and operate over cache instances. */ +@SuppressWarnings("nullness") +public final class Caches { + + /** A cache that never stores any values. */ + public static <K, V> Cache<K, V> noop() { + return new SubCacheable<K, V>() { + @Override + public V peek(K key) { + return null; + } + + @Override + public V computeIfAbsent(K key, Function<K, V> loadingFunction) { + return loadingFunction.apply(key); + } + + @Override + public void put(K key, V value) {} + + @Override + public void clear() {} + + @Override + public void remove(K key) {} + + @Override + public Set<K> keys() { + return Collections.emptySet(); + } + }; + } + + /** + * Uses the specified {@link PipelineOptions} to configure and return a cache instance based upon + * parameters within {@link SdkHarnessOptions}. + */ + public static <K, V> Cache<K, V> fromOptions(PipelineOptions options) { + // We specifically use cache2k since it allows for recursive computeIfAbsent calls + // preventing deadlock from occurring when a loading function mutates the underlying cache + org.cache2k.Cache<Object, Object> cache = + Cache2kBuilder.forUnknownTypes() + .maximumWeight( + options.as(SdkHarnessOptions.class).getMaxCacheMemoryUsageMb() * 1024L * 1024L) + .weigher( + new Weigher<Object, Object>() { + private final MemoryMeter memoryMeter = MemoryMeter.builder().build(); + + @Override + public int weigh(Object key, Object value) { + long size = memoryMeter.measureDeep(key) + memoryMeter.measureDeep(value); + return size > Integer.MAX_VALUE ? Integer.MAX_VALUE : (int) size; + } + }) + .storeByReference(true) + .executor(MoreExecutors.directExecutor()) + .build(); + + return (Cache<K, V>) forCache(cache); + } + + /** + * Returns a view of a cache that operates on keys with a specified key prefix. + * + * <p>All lookups, insertions, and removals into the parent {@link Cache} will be prefixed by the + * specified prefixes. + * + * <p>Operations which operate over the entire caches contents such as {@link Cache#clear} only + * operate over keys with the specified prefixes. + */ + public static <K, V> Cache<K, V> subCache( + Cache<?, ?> cache, Object keyPrefix, Object... additionalKeyPrefix) { + if (cache instanceof SubCache) { + return new SubCache<>( + ((SubCache<?, ?>) cache).cache, + ((SubCache<?, ?>) cache).keyPrefix.subKey(keyPrefix, additionalKeyPrefix)); + } else if (cache instanceof SubCacheable) { + return new SubCache<>((SubCacheable<?, ?>) cache, CompositeKey.ROOT.subKey(keyPrefix)); + } + throw new UnsupportedOperationException("An unsupported type of cache was passed."); + } + + /** A cache that never evicts any values. */ + public static <K, V> Cache<K, V> eternal() { + // We specifically use cache2k since it allows for recursive computeIfAbsent calls Review comment: Yes, `computeIfAbsent` creating a loop is definitely an issue that could happen if this API was misused. I have needed a chain for the use case that I'm building like X -> Y -> Z and is guaranteed to not create a loop otherwise it will cause an infinite loop outside of the cache as well since I'm effectively caching pages and continuation tokens. This is likely to change though as a different implementation won't have the need for the recursive `computeIfAbsent`. I'll update the docs in the future (and tests) whether this is needed or not depending on the different implementation. On a different note, I'm trying to ensure that the cache reduces the number of OOMs and maximizes the amount of cache hits by having one cache that is used and a bunch of sub caches that provide views over them for different key spaces. Our previous implementation used distinct cache objects which required one to partition memory and give each cache a limited amount of memory. There didn't seem to be a good way to have the caches work together so that if one was unused the other could have more memory. Note that all of these caches used only weight based eviction. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
