refactor: split get clusters and print cluster examples into seperate functions

Author: abhitrueprogrammer

ConsoleApp2/Analyser.cs

@@ -119,13 +119,18 @@ public static ITransformer TrainKMeansClusterer(MLContext mlContext, IDataView t
             return pipeline.Fit(trainData);
         }
 
-        public static void PrintClusterExamples(MLContext mlContext, IDataView data, ITransformer model)
+        public static IEnumerable<IGrouping<uint, CommitPredictionWithData>> GetClusters(MLContext mlContext, IDataView data, ITransformer model)
         {
             var predictions = model.Transform(data);
             var results = mlContext.Data.CreateEnumerable<CommitPredictionWithData>(predictions, reuseRowObject: false).ToList();
-
-            var clusters = results.GroupBy(x => x.PredictedClusterId).OrderBy(g => g.Key);
-
+            return results.GroupBy(x => x.PredictedClusterId).OrderBy(g => g.Key);
+        }
+        public static IEnumerable<IGrouping<uint, CommitPredictionWithData>> PredictClusterNames(IEnumerable<IGrouping<uint, CommitPredictionWithData>> clusters)
+        {
+            return null;
+        }
+        public static void PrintClusterExamples(IEnumerable<IGrouping<uint, CommitPredictionWithData>> clusters)
+        {
             Console.WriteLine("\n--- Cluster Examples ---");
             foreach (var cluster in clusters)
             {

ConsoleApp2/Program.cs

@@ -41,7 +41,8 @@ static async Task Main(string[] args)
                 var model = Analyser.TrainKMeansClusterer(mlContext, split.TrainSet, featurizer, bestK);
 
                 // 6. Print 2 examples from each cluster
-                Analyser.PrintClusterExamples(mlContext, split.TrainSet, model);
+                var clusters = Analyser.GetClusters(mlContext, split.TrainSet, model);
+                Analyser.PrintClusterExamples(clusters);
             }
         }
     }

ConsoleApp2/TODO.md

@@ -1,4 +1,5 @@
 1. Add vector embeddings instead of featurize text
 2. Try out DB scan 
 3. Try Kmeans visualisation
-4. Label clusters using LLM
+4. Label clusters using LLM.
+5. Find a way to add lables to the trained model so that we can use it for classification as well.