Port sequence confidence changes

src/SIL.Machine/QualityEstimation/ChrF3QualityEstimator.cs

@@ -2,6 +2,7 @@
 using System.Collections.Generic;
 using System.Linq;
 using SIL.Machine.Corpora;
+using SIL.Machine.Statistics;
 
 namespace SIL.Machine.QualityEstimation
 {
@@ -83,22 +84,6 @@ ScriptureBookScores bookScores
             return ComputeSegmentUsability(segmentScores, chapterScores, bookScores);
         }
 
-        /// <summary>
-        /// Calculates the geometric mean for a collection of values.
-        /// </summary>
-        /// <param name="values"></param>
-        /// <returns>The geometric mean.</returns>
-        private static double GeometricMean(IList<double> values)
-        {
-            // Geometric mean requires positive values
-            if (values == null || !values.Any() || values.Any(x => x <= 0))
-                return 0;
-
-            // Compute the sum of the natural logarithms of all values,
-            // and divide by the count of numbers and take the exponential
-            return Math.Exp(values.Sum(Math.Log) / values.Count);
-        }
-
         private double CalculateUsableProbability(double chrF3)
         {
             double usableWeight = Math.Exp(-Math.Pow(chrF3 - Usable.Mean, 2) / (2 * Usable.Variance)) * Usable.Count;
@@ -267,7 +252,7 @@ private List<TextUsability> ComputeTextUsability(TextScores textScores)
             {
                 textScores.AddScore(
                     textIdConfidences.Key,
-                    new Score(_slope, confidence: GeometricMean(textIdConfidences.Value), _intercept)
+                    new Score(_slope, confidence: StatisticalMethods.GeometricMean(textIdConfidences.Value), _intercept)
                 );
             }
 
@@ -325,7 +310,11 @@ out List<double> bookAndChapterConfidences
                 chapterScores.AddScore(
                     bookAndChapterConfidences.Key.Book,
                     bookAndChapterConfidences.Key.Chapter,
-                    new Score(_slope, confidence: GeometricMean(bookAndChapterConfidences.Value), _intercept)
+                    new Score(
+                        _slope,
+                        confidence: StatisticalMethods.GeometricMean(bookAndChapterConfidences.Value),
+                        _intercept
+                    )
                 );
             }
 
@@ -334,7 +323,7 @@ out List<double> bookAndChapterConfidences
             {
                 bookScores.AddScore(
                     bookConfidences.Key,
-                    new Score(_slope, confidence: GeometricMean(bookConfidences.Value), _intercept)
+                    new Score(_slope, confidence: StatisticalMethods.GeometricMean(bookConfidences.Value), _intercept)
                 );
             }
 

src/SIL.Machine/Statistics/StatisticalMethods.cs

@@ -43,5 +43,21 @@ public static double KullbackLeiblerDivergence(IEnumerable<double> dist1, IEnume
         {
             return dist1.Zip(dist2, (p1, p2) => p1 == 0 || p2 == 0 ? 0 : Math.Log(p1 / p2, 2) * p1).Sum();
         }
+
+        /// <summary>
+        /// Calculates the geometric mean for a collection of values.
+        /// </summary>
+        /// <param name="values"></param>
+        /// <returns>The geometric mean.</returns>
+        public static double GeometricMean(IList<double> values)
+        {
+            // Geometric mean requires positive values
+            if (values == null || !values.Any() || values.Any(x => x <= 0))
+                return 0;
+
+            // Compute the sum of the natural logarithms of all values,
+            // and divide by the count of numbers and take the exponential
+            return Math.Exp(values.Sum(Math.Log) / values.Count);
+        }
     }
 }

src/SIL.Machine/Translation/HybridTranslationEngine.cs

@@ -3,6 +3,7 @@
 using System.Linq;
 using System.Threading;
 using System.Threading.Tasks;
+using SIL.Machine.Statistics;
 using SIL.Machine.Tokenization;
 using SIL.ObjectModel;
 
@@ -488,6 +489,7 @@ private TranslationResult Merge(TranslationResult interactiveResult, Translation
                 interactiveResult.SourceTokens,
                 mergedTargetSegment,
                 mergedConfidences,
+                StatisticalMethods.GeometricMean(mergedConfidences),
                 mergedSources,
                 alignment,
                 interactiveResult.Phrases

src/SIL.Machine/Translation/TransferEngine.cs

@@ -5,6 +5,7 @@
 using SIL.Machine.Annotations;
 using SIL.Machine.Corpora;
 using SIL.Machine.Morphology;
+using SIL.Machine.Statistics;
 using SIL.Machine.Tokenization;
 using SIL.ObjectModel;
 
@@ -186,6 +187,7 @@ public IReadOnlyList<TranslationResult> Translate(int n, IReadOnlyList<string> s
                         segment,
                         targetTokens,
                         confidences,
+                        StatisticalMethods.GeometricMean(confidences),
                         sources,
                         alignment,
                         new[] { new Phrase(Range<int>.Create(0, normalizedSourceTokens.Count), targetWords.Count) }

src/SIL.Machine/Translation/TranslationExtensions.cs

@@ -120,6 +120,7 @@ public static TranslationResult Truecase(
                 result.SourceTokens,
                 targetTokens,
                 result.Confidences,
+                result.SequenceConfidence,
                 result.Sources,
                 result.Alignment,
                 result.Phrases

src/SIL.Machine/Translation/TranslationResult.cs

@@ -11,6 +11,7 @@ public TranslationResult(
             IEnumerable<string> sourceTokens,
             IEnumerable<string> targetTokens,
             IEnumerable<double> confidences,
+            double sequenceConfidence,
             IEnumerable<TranslationSources> sources,
             WordAlignmentMatrix alignment,
             IEnumerable<Phrase> phrases
@@ -27,6 +28,7 @@ IEnumerable<Phrase> phrases
                     nameof(confidences)
                 );
             }
+            SequenceConfidence = sequenceConfidence;
             Sources = sources.ToArray();
             if (Sources.Count != TargetTokens.Count)
             {
@@ -58,6 +60,7 @@ IEnumerable<Phrase> phrases
         public IReadOnlyList<string> SourceTokens { get; }
         public IReadOnlyList<string> TargetTokens { get; }
         public IReadOnlyList<double> Confidences { get; }
+        public double SequenceConfidence { get; }
         public IReadOnlyList<TranslationSources> Sources { get; }
         public WordAlignmentMatrix Alignment { get; }
         public IReadOnlyList<Phrase> Phrases { get; }

src/SIL.Machine/Translation/TranslationResultBuilder.cs

@@ -11,6 +11,7 @@ public class TranslationResultBuilder
         private readonly List<double> _confidences;
         private readonly List<TranslationSources> _sources;
         private readonly List<PhraseInfo> _phrases;
+        private readonly double _sequenceConfidences;
 
         public TranslationResultBuilder(IReadOnlyList<string> sourceTokens)
         {
@@ -19,6 +20,7 @@ public TranslationResultBuilder(IReadOnlyList<string> sourceTokens)
             _confidences = new List<double>();
             _sources = new List<TranslationSources>();
             _phrases = new List<PhraseInfo>();
+            _sequenceConfidences = -1.0;
         }
 
         public IReadOnlyList<string> SourceTokens { get; }
@@ -29,6 +31,7 @@ public TranslationResultBuilder(IReadOnlyList<string> sourceTokens)
         public IReadOnlyList<double> Confidences => _confidences;
         public IReadOnlyList<TranslationSources> Sources => _sources;
         public IReadOnlyList<PhraseInfo> Phrases => _phrases;
+        public double SequenceConfidences => _sequenceConfidences;
 
         public void AppendToken(string token, TranslationSources source, double confidence)
         {
@@ -246,6 +249,7 @@ public TranslationResult ToResult(string translation = null)
                 SourceTokens,
                 _targetTokens,
                 _confidences,
+                _sequenceConfidences,
                 sources,
                 alignment,
                 phrases