/*
    * 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
    *
    *      https://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.
   */
  
  /*
   * This is not the original file distributed by the Apache Software Foundation
   * It has been modified by the Hipparchus project
   */
  package org.hipparchus.stat.inference;
  
  import org.hipparchus.exception.MathIllegalArgumentException;
  import org.hipparchus.exception.NullArgumentException;
  import org.junit.Assert;
  import org.junit.Test;
  
  /**
   * Test cases for the MannWhitneyUTest class.
   */
  
  public class MannWhitneyUTestTest {
  
      protected MannWhitneyUTest testStatistic = new MannWhitneyUTest();
  
      /**
       * Target values for most tests below were computed using R, version 3.4.4.
       * The wilcox.test function was used throughout, passing (along with x and y)
       *   alternative = "two.sized"
       *   mu = 0
       *   paired = FALSE
       * For the exact tests,
       *   exact = TRUE
       * For normal approximation (exact = false in MannWhitneyTest API)
       *   exact = FALSE
       *   correct = TRUE (tells R to do the continuity correction).
       * The target for the U statistic value was obtained by running
       * the R test twice, the second time switching the order of x and y and
       * taking the minimum value. This is correct because the statistic that R
       * reports as W is the Wilcoxon Rank Sum with x before y and U is the
       * minimum of this value and the rank sum with y before x.
       */
  
      @Test
      public void testMannWhitneyUSimple() {
          /*
           * Target values computed using R version 3.4.4.
           */
          final double[] x = {
              19, 22, 16, 29, 24
          };
          final double[] y = {
              20, 11, 17, 12
          };
  
          Assert.assertEquals(3, testStatistic.mannWhitneyU(x, y), 1e-10);
          Assert.assertEquals(0.11134688653,
                              testStatistic.mannWhitneyUTest(x, y, false), 1e-5);
          Assert.assertEquals(0.11111111111,
                              testStatistic.mannWhitneyUTest(x, y, true), 1e-5);
      }
  
      @Test
      public void testDisjoint() {
          /*
           * Target values computed using R version 3.4.4.
           */
          final double[] x = {
              1, 2, 3, 4, 5
          };
          final double[] y = {
              6, 7, 8, 9, 10, 11
          };
          Assert.assertEquals(0, testStatistic.mannWhitneyU(x, y), 1e-10);
          Assert.assertEquals(0.0081131172656,
                              testStatistic.mannWhitneyUTest(x, y, false), 1e-5);
          Assert.assertEquals(0.004329004329,
                              testStatistic.mannWhitneyUTest(x, y, true), 1e-5);
  
      }
  
      @Test
      public void testMannWhitneyUInputValidation() {
          /*
           * Samples must be present, i.e. length > 0
           */
          try {
              testStatistic.mannWhitneyUTest(new double[] {}, new double[] {
                 1.0
             });
             Assert
                 .fail("x does not contain samples (exact), MathIllegalArgumentException expected");
         } catch (MathIllegalArgumentException ex) {
             // expected
         }
 
         try {
             testStatistic.mannWhitneyUTest(new double[] {
                 1.0
             }, new double[] {});
             Assert
                 .fail("y does not contain samples (exact), MathIllegalArgumentException expected");
         } catch (MathIllegalArgumentException ex) {
             // expected
         }
 
         /*
          * x and y is null
          */
         try {
             testStatistic.mannWhitneyUTest(null, null);
             Assert
                 .fail("x and y is null (exact), NullArgumentException expected");
         } catch (NullArgumentException ex) {
             // expected
         }
 
         try {
             testStatistic.mannWhitneyUTest(null, null);
             Assert
                 .fail("x and y is null (asymptotic), NullArgumentException expected");
         } catch (NullArgumentException ex) {
             // expected
         }
 
         /*
          * x or y is null
          */
         try {
             testStatistic.mannWhitneyUTest(null, new double[] {
                 1.0
             });
             Assert.fail("x is null (exact), NullArgumentException expected");
         } catch (NullArgumentException ex) {
             // expected
         }
 
         try {
             testStatistic.mannWhitneyUTest(new double[] {
                 1.0
             }, null);
             Assert.fail("y is null (exact), NullArgumentException expected");
         } catch (NullArgumentException ex) {
             // expected
         }
     }
 
     @Test
     public void testLargeDatasetExact() {
         /**
          * Expected values computed using R 3.4.4
          */
         final double[] x = {
             11, 22, 19, 22.3, 16, 29, 24, 5.2, 7, 3, 44, 72, 43, 18, 65
         };
         final double[] y = {
             15, 32, 38, 5, 6, 29.1, 31, 73, 88, 70, 50, 60, 93, 112, 190
         };
 
         Assert.assertEquals(59, testStatistic.mannWhitneyU(x, y), 1e-10);
         Assert.assertEquals(0.027925057353,
                             testStatistic.mannWhitneyUTest(x, y, false), 1e-9);
         Assert.assertEquals(0.02635404434,
                             testStatistic.mannWhitneyUTest(x, y, true), 1e-9);
         // Should default to exact test
         Assert.assertEquals(0.02635404434, testStatistic.mannWhitneyUTest(x, y),
                             1e-9);
     }
 
     @Test
     public void testDatasetTooLargeForExact() {
         /**
          * Expected values computed using R 3.4.4
          */
         final double[] x = {
             11, 22, 19, 22.3, 16, 29, 24, 5.2, 7, 3, 44, 72, 43, 18, 65, 69, 71,
             115, 117, 119, 121, 123, 124, 125, 126, 127
         };
         final double[] y = {
             15, 32, 38, 5, 6, 29.1, 31, 73, 88, 70, 50, 60, 93, 112, 190, 200,
             201, 202, 203, 204, 205, 207, 209, 210, 212
         };
 
         Assert.assertEquals(204, testStatistic.mannWhitneyU(x, y), 1e-10);
         Assert.assertEquals(0.023177956065,
                             testStatistic.mannWhitneyUTest(x, y, false), 1e-9);
         // Should default to normal approximation
         Assert.assertEquals(0.023177956065,
                             testStatistic.mannWhitneyUTest(x, y), 1e-9);
         Assert.assertEquals(0.022259264963,
                             testStatistic.mannWhitneyUTest(x, y, true), 1e-9);
     }
 
     @Test
     public void testExactHiP() {
         /**
          * Expected values computed using R 3.4.4
          */
         final double[] x = {
             0, 2, 4, 6, 8, 10, 12, 14, 16, 18
         };
         final double[] y = {
             1, 3, 5, 7, 9, 11, 13, 15, 17, 19
         };
         Assert.assertEquals(45, testStatistic.mannWhitneyU(x, y), 1e-10);
         Assert.assertEquals(0.7337299957,
                             testStatistic.mannWhitneyUTest(x, y, false), 1e-9);
         Assert.assertEquals(0.73936435082,
                             testStatistic.mannWhitneyUTest(x, y, true), 1e-9);
 
     }
 
     @Test
     public void testExactLarge() {
         /**
          * Expected values computed using R 3.4.4
          */
         final double[] x = {
             0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30, 32, 34,
             36, 38, 40, 42, 44, 46, 48, 50, 52, 54, 56, 58, 60
         };
         final double[] y = {
             1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29, 31, 33, 35,
             37, 39, 43, 45, 49, 51, 55, 61, 63, 65, 67, 69
         };
         Assert.assertEquals(441, testStatistic.mannWhitneyU(x, y), 1e-10);
         Assert.assertEquals(0.73459710599,
                             testStatistic.mannWhitneyUTest(x, y, false), 1e-9);
         Assert.assertEquals(0.73642668965,
                             testStatistic.mannWhitneyUTest(x, y, true), 1e-9);
     }
 
     @Test
     public void testExactVerySmall() {
         /**
          * Expected values computed using R 3.4.4
          */
         final double[] x = {
             1,2
         };
         final double[] y = {
             1.5, 2.5
         };
         Assert.assertEquals(1, testStatistic.mannWhitneyU(x, y), 1e-10);
         Assert.assertEquals(0.66666666667,
                             testStatistic.mannWhitneyUTest(x, y, true), 1e-9);
         Assert.assertEquals(0.6985353583,
                             testStatistic.mannWhitneyUTest(x, y, false), 1e-9);
 
     }
 
     @Test
     public void testExactDegenerate() {
         final double[] x = {
             1
         };
         final double[] y = {
             1.5
         };
         Assert.assertEquals(0, testStatistic.mannWhitneyU(x, y), 1e-10);
         Assert.assertEquals(1.0,
                             testStatistic.mannWhitneyUTest(x, y, true), 1e-9);
         Assert.assertEquals(1.0,
                             testStatistic.mannWhitneyUTest(x, y, false), 1e-9);
     }
 
     @Test
     public void testApproximateWithTies() {
         /**
          * Expected values computed using R 3.4.4
          */
         final double[] x = {
             0, 2, 4, 6, 8, 10, 12, 14, 16, 18
         };
         final double[] y = {
             1, 3, 5, 7, 8, 10, 10, 13, 15, 17, 19
         };
         try {
             testStatistic.mannWhitneyUTest(x, y, true);
         } catch (MathIllegalArgumentException ex) {
             // expected
         }
         Assert.assertEquals(50.5, testStatistic.mannWhitneyU(x, y), 1e-10);
         Assert.assertEquals(0.77784391371,
                             testStatistic.mannWhitneyUTest(x, y, false), 1e-9);
 
     }
 
     @Test
     public void testApproximateWithTies2() {
         /**
          * Expected values computed using R 3.4.4
          */
         final double[] x = {
             2, 3, 7, 11, 23, 45, 48, 55, 70, 81, 92, 95, 97, 100, 110, 123, 125
         };
         final double[] y = {
             3.5, 4, 8, 12, 25, 46, 49, 56, 70, 81, 92, 95, 97, 97, 100, 112,
             125, 127
         };
         try {
             testStatistic.mannWhitneyUTest(x, y, true);
         } catch (MathIllegalArgumentException ex) {
             // expected
         }
         Assert.assertEquals(142, testStatistic.mannWhitneyU(x, y), 1e-10);
         Assert.assertEquals(0.72874565042,
                             testStatistic.mannWhitneyUTest(x, y, false), 1e-9);
 
     }
 
     @Test
     public void testIdenticalArrays() {
         /**
          * Expected values computed using R 3.4.4
          */
         final double[] x = {
             1, 2, 3, 4, 5, 6, 7, 8, 9, 10
         };
         final double[] y = {
             1, 2, 3, 4, 5, 6, 7, 8, 9, 10
         };
         try {
             testStatistic.mannWhitneyUTest(x, y, true);
         } catch (MathIllegalArgumentException ex) {
             // expected
         }
         Assert.assertEquals(50, testStatistic.mannWhitneyU(x, y), 1e-10);
         Assert.assertEquals(1.0, testStatistic.mannWhitneyUTest(x, y, false),
                             1e-10);
 
     }
 
     @Test(expected = MathIllegalArgumentException.class)
     public void testExactThrowsOnTies() {
         final double[] x = {
             1, 5, 7
         };
         final double[] y = {
             2, 3, 1
         };
         testStatistic.mannWhitneyUTest(x, y, true);
     }
 
     @Test
     public void testBigDataSet() {
         double[] d1 = new double[1500];
         double[] d2 = new double[1500];
         for (int i = 0; i < 1500; i++) {
             d1[i] = 2 * i;
             d2[i] = 2 * i + 1;
         }
         double result = testStatistic.mannWhitneyUTest(d1, d2);
         Assert.assertTrue(result > 0.1);
     }
 
     @Test
     public void testBigDataSetOverflow() {
         // MATH-1145
         double[] d1 = new double[110000];
         double[] d2 = new double[110000];
         for (int i = 0; i < 110000; i++) {
             d1[i] = i;
             d2[i] = i;
         }
         Assert.assertEquals(1.0, testStatistic.mannWhitneyUTest(d1, d2, false),
                             1E-7);
     }
 }