Merge pull request #349 from NOAA-OWP/issue342

Fix a bug in determining the minimum sample size for sampling uncerta…
NOAA-OWP · Oct 28, 2024 · fbe4fb1 · fbe4fb1
2 parents 769e5c5 + b5bc780
commit fbe4fb1
Show file tree

Hide file tree

Showing 2 changed files with 15 additions and 10 deletions.
diff --git a/wres-datamodel/src/wres/datamodel/bootstrap/BootstrapUtilities.java b/wres-datamodel/src/wres/datamodel/bootstrap/BootstrapUtilities.java
@@ -47,6 +47,7 @@ public class BootstrapUtilities
      * @param <R> the type of right-ish time-series data
      * @param pool the pool
      * @return the optimal block size for the stationary bootstrap in timestep units, together with the timestep
+     * @throws InsufficientDataForResamplingException if there is insufficient data to calculate the optimal block size
      */
     public static <R> Pair<Long, Duration> getOptimalBlockSizeForStationaryBootstrap( Pool<TimeSeries<Pair<Double, R>>> pool )
     {
@@ -110,7 +111,7 @@ public static <R> Pair<Long, Duration> getOptimalBlockSizeForStationaryBootstrap
 
     /**
      * Determines whether sufficient data is available for bootstrap resampling. There must be more than one event
-     * across time time-series present.
+     * across the consolidated time-series present.
      * @param <T> the type of time-series data
      * @param data the time-series data
      * @return whether there is sufficient data for the stationary bootstrap
@@ -120,10 +121,13 @@ public static <T> boolean hasSufficientDataForStationaryBootstrap( List<TimeSeri
     {
         Objects.requireNonNull( data );
 
+        // Consolidate the time-series by event datetime and count the number of events
         long eventCount = data.stream()
-                              .mapToLong( t -> t.getEvents()
-                                                .size() )
-                              .sum();
+                              .flatMap( t -> t.getEvents()
+                                              .stream()
+                                              .map( Event::getTime ) )
+                              .distinct()
+                              .count();
 
         if ( LOGGER.isDebugEnabled() )
         {
@@ -413,6 +417,7 @@ private static void addQuantileForBoxplot( List<BoxplotStatistic> quantile,
      * @param <T> the type of time-series data
      * @param pool the pool
      * @return the optimal block size for the stationary bootstrap
+     * @throws InsufficientDataForResamplingException if there are fewer than two events across the consolidated series
      */
     private static <T> Pair<Long, Duration> getOptimalBlockSizeForStationaryBootstrap( List<TimeSeries<Pair<Double, T>>> pool )
     {

diff --git a/wres-datamodel/test/wres/datamodel/time/TimeSeriesTest.java b/wres-datamodel/test/wres/datamodel/time/TimeSeriesTest.java
@@ -26,7 +26,7 @@
 
 /**
  * Tests the {@link TimeSeries}
- * 
+ *
  * @author James Brown
  */
 
@@ -79,10 +79,10 @@ public void runBeforeEachTest()
                                        FEATURE_NAME,
                                        UNIT );
         this.testSeries = builder.setMetadata( metadata )
-                            .addEvent( iterator.next() )
-                            .addEvent( iterator.next() )
-                            .addEvent( iterator.next() )
-                            .build();
+                                 .addEvent( iterator.next() )
+                                 .addEvent( iterator.next() )
+                                 .addEvent( iterator.next() )
+                                 .build();
     }
 
     /**
@@ -146,7 +146,7 @@ public void testHashCode()
     @Test
     public void testEquals()
     {
-        // Reflexive 
+        // Reflexive
         assertEquals( this.testSeries, this.testSeries );
 
         // Symmetric