ishnagy commented on code in PR #50933: URL: https://github.com/apache/spark/pull/50933#discussion_r2095965583
########## common/sketch/src/test/java/org/apache/spark/util/sketch/TestSparkBloomFilter.java: ########## @@ -0,0 +1,294 @@ +/* + * 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.spark.util.sketch; + +import org.junit.jupiter.api.*; +import org.junitpioneer.jupiter.cartesian.CartesianTest; +import org.junitpioneer.jupiter.cartesian.CartesianTest.Values; + +import java.time.Duration; +import java.time.Instant; +import java.util.Random; + +@Disabled +public class TestSparkBloomFilter { + + // the implemented fpp limit is only approximating the hard boundary, + // so we'll need an error threshold for the assertion + final double FPP_EVEN_ODD_ERROR_FACTOR = 0.05; + final double FPP_RANDOM_ERROR_FACTOR = 0.04; + + final long ONE_GB = 1024L * 1024L * 1024L; + final long REQUIRED_HEAP_UPPER_BOUND_IN_BYTES = 4 * ONE_GB; + + private static Instant START; + private Instant start; + + @BeforeAll + public static void beforeAll() { + START = Instant.now(); + } + + @AfterAll + public static void afterAll() { + Duration duration = Duration.between(START, Instant.now()); + System.err.println(duration + " TOTAL"); + } + + @BeforeEach + public void beforeEach() { + start = Instant.now(); + } + + @AfterEach + public void afterEach(TestInfo testInfo) { + Duration duration = Duration.between(start, Instant.now()); + System.err.println(duration + " " + testInfo.getDisplayName()); + } + + /** + * This test, in N number of iterations, inserts N even numbers (2*i) int, + * and leaves out N odd numbers (2*i+1) from the tested BloomFilter instance. + * + * It checks the 100% accuracy of mightContain=true on all of the even items, + * and measures the mightContain=true (false positive) rate on the not-inserted odd numbers. + * + * @param numItems the number of items to be inserted + * @param expectedFpp the expected fpp rate of the tested BloomFilter instance + * @param deterministicSeed the deterministic seed to use to initialize + * the primary BloomFilter instance. + */ + @CartesianTest + public void testAccuracyEvenOdd( + @Values(longs = {1_000_000L, 1_000_000_000L, 5_000_000_000L}) long numItems, + @Values(doubles = {0.05, 0.03, 0.01, 0.001}) double expectedFpp, + @Values(ints = {BloomFilterImpl.DEFAULT_SEED, 1, 127}) int deterministicSeed + ) { + long optimalNumOfBits = BloomFilter.optimalNumOfBits(numItems, expectedFpp) / Byte.SIZE; + System.err.printf( + "optimal bitArray: %d (%d MB)\n", + optimalNumOfBits, + optimalNumOfBits / Byte.SIZE / 1024 / 1024 + ); + Assumptions.assumeTrue( + optimalNumOfBits / Byte.SIZE < REQUIRED_HEAP_UPPER_BOUND_IN_BYTES, + "this testcase would require allocating more than 4GB of heap mem (" + + optimalNumOfBits + + " bits)" + ); + + BloomFilter bloomFilter = BloomFilter.create(numItems, optimalNumOfBits, deterministicSeed); + System.err.printf( + "allocated bitArray: %d (%d MB)\n", + bloomFilter.bitSize(), + bloomFilter.bitSize() / Byte.SIZE / 1024 / 1024 + ); + + for (long i = 0; i < numItems; i++) { + if (i % 10_000_000 == 0) { + System.err.printf( + "i: %d, bitCount: %d, saturation: %f\n", + i, + bloomFilter.cardinality(), + (double) bloomFilter.cardinality() / bloomFilter.bitSize() + ); + } + bloomFilter.putLong(2 * i); + } + + long mightContainEven = 0; + long mightContainOdd = 0; + + for (long i = 0; i < numItems; i++) { + if (i % (numItems / 100) == 0) { + System.err.print("."); + System.err.flush(); + } + + long even = 2 * i; + if (bloomFilter.mightContainLong(even)) { + mightContainEven++; + } + + long odd = 2 * i + 1; + if (bloomFilter.mightContainLong(odd)) { + mightContainOdd++; + } + } + System.err.println(); + + Assertions.assertEquals( + numItems, mightContainEven, + "mightContainLong must return true for all inserted numbers" + ); + + double actualFpp = (double) mightContainOdd / numItems; + double acceptableFpp = expectedFpp * (1 + FPP_EVEN_ODD_ERROR_FACTOR); + + System.err.printf("expectedFpp: %f %%\n", 100 * expectedFpp); + System.err.printf("acceptableFpp: %f %%\n", 100 * acceptableFpp); + System.err.printf("actualFpp: %f %%\n", 100 * actualFpp); + + Assumptions.assumeTrue( + actualFpp <= acceptableFpp, + String.format( + "acceptableFpp(%f %%) < actualFpp (%f %%)", + 100 * acceptableFpp, + 100 * actualFpp + ) + ); + + Assertions.assertTrue( + actualFpp <= acceptableFpp, + String.format( + "acceptableFpp(%f %%) < actualFpp (%f %%)", + 100 * acceptableFpp, + 100 * actualFpp + ) + ); + } + + /** + * This test inserts N pseudorandomly generated numbers in 2N number of iterations in two + * differently seeded (theoretically independent) BloomFilter instances. All the random + * numbers generated in an even-iteration will be inserted into both filters, all the + * random numbers generated in an odd-iteration will be left out from both. + * + * The test checks the 100% accuracy of 'mightContain=true' for all the items inserted + * in an even-loop. It counts the false positives as the number of odd-loop items for + * which the primary filter reports 'mightContain=true', but secondary reports + * 'mightContain=false'. Since we inserted the same elements into both instances, + * and the secondary reports non-insertion, the 'mightContain=true' from the primary + * can only be a false positive. + * + * @param numItems the number of items to be inserted + * @param expectedFpp the expected fpp rate of the tested BloomFilter instance + * @param deterministicSeed the deterministic seed to use to initialize + * the primary BloomFilter instance. (The secondary will be + * initialized with the constant seed of 0xCAFEBABE) + */ + @CartesianTest + public void testAccuracyRandom( + @Values(longs = {1_000_000L, 1_000_000_000L}) long numItems, + @Values(doubles = {0.05, 0.03, 0.01, 0.001}) double expectedFpp, + @Values(ints = {BloomFilterImpl.DEFAULT_SEED, 1, 127}) int deterministicSeed + ) { + long optimalNumOfBits = BloomFilter.optimalNumOfBits(numItems, expectedFpp); + System.err.printf( + "optimal bitArray: %d (%d MB)\n", + optimalNumOfBits, + optimalNumOfBits / Byte.SIZE / 1024 / 1024 + ); + Assumptions.assumeTrue( + 2 * optimalNumOfBits / Byte.SIZE < REQUIRED_HEAP_UPPER_BOUND_IN_BYTES, + "this testcase would require allocating more than 4GB of heap mem (2x " + + optimalNumOfBits + + " bits)" + ); + + BloomFilter bloomFilterPrimary = + BloomFilter.create(numItems, optimalNumOfBits, deterministicSeed); + BloomFilter bloomFilterSecondary = + BloomFilter.create(numItems, optimalNumOfBits, 0xCAFEBABE); + + System.err.printf( + "allocated bitArray: %d (%d MB)\n", + bloomFilterPrimary.bitSize(), + bloomFilterPrimary.bitSize() / Byte.SIZE / 1024 / 1024 + ); + + + Random pseudoRandom = new Random(); + long iterationCount = 2 * numItems; + + pseudoRandom.setSeed(deterministicSeed); + for (long i = 0; i < iterationCount; i++) { + if (i % 10_000_000 == 0) { + System.err.printf( + "i: %d, bitCount: %d, saturation: %f\n", + i, + bloomFilterPrimary.cardinality(), + (double) bloomFilterPrimary.cardinality() / bloomFilterPrimary.bitSize() + ); + } + + long candidate = pseudoRandom.nextLong(); + if (i % 2 == 0) { + bloomFilterPrimary.putLong(candidate); + bloomFilterSecondary.putLong(candidate); + } + } + + long mightContainEvenIndexed = 0; + long mightContainOddIndexed = 0; + + pseudoRandom.setSeed(deterministicSeed); + for (long i = 0; i < iterationCount; i++) { + if (i % (iterationCount / 100) == 0) { + System.err.print("."); + System.err.flush(); + } + + long candidate = pseudoRandom.nextLong(); + if (bloomFilterPrimary.mightContainLong(candidate)) { + if (i % 2 == 0) { + mightContainEvenIndexed++; + } else { + // only count those cases as false positives, + // where the secondary has confirmed, + // that we haven't inserted before + // (mitigating duplicates in input sequence) + if (!bloomFilterSecondary.mightContainLong(candidate)) { + mightContainOddIndexed++; + } + } + } + } + System.err.println(); + + Assertions.assertEquals( + numItems, mightContainEvenIndexed, + "mightContainLong must return true for all inserted numbers" + ); + + double actualFpp = (double) mightContainOddIndexed / numItems; Review Comment: indeed, it should probably be very close to the proper value, but this calculation doesn't account for the odd indexes ignored based on the secondary's result. let me try to address that somehow. -- 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: reviews-unsubscr...@spark.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
