Added score-based loop

GlassOfWhiskey · GlassOfWhiskey · commit 6ae19de32cf9 · 2025-03-26T00:17:35.000+01:00
diff --git a/Makefile b/Makefile
@@ -37,10 +37,10 @@ $(OBJ_DIR)/common.o: $(SRC_DIR)/common.cpp
 	$(CXX) $(CXXFLAGS) -c $< -o $@
 
 $(OBJ_DIR)/kmeans_cl.o: $(KMEANS_LIB_DIR)/kmeans_cl.c $(KMEANS_LIB_DIR)/kmeans_cl.h
-	$(CC) $(CFLAGS) -c $< -o $@
+	$(CXX) $(CXXFLAGS) -c $< -o $@
 
 $(OBJ_DIR)/kmeans.o: $(KMEANS_LIB_DIR)/kmeans.c
-	$(CC) $(CFLAGS) -c $< -o $@
+	$(CXX) $(CXXFLAGS) -c $< -o $@
 
 $(OBJ_DIR)/dbscan.o: $(DBSCAN_LIB_DIR)/dbscan.cpp $(DBSCAN_LIB_DIR)/dbscan.hpp
 	$(CXX) $(CXXFLAGS) -c $< -o $@
diff --git a/SimulatedAnnealing b/SimulatedAnnealing
@@ -1 +1 @@
-Subproject commit 7a152839831513fa1e83deb92ccfbe1f448ed47c
+Subproject commit 61f0d71f35ca3540de0a0229718c7c4cbe9b5e90
diff --git a/common.cpp b/common.cpp
@@ -1,4 +1,5 @@
 #include "common.h"
+#include <algorithm>
 #include <cmath>
 #include <cstddef> // For offsetof
 #include <fstream>
diff --git a/vendor/kmeans/kmeans.c b/vendor/kmeans/kmeans.c
@@ -68,17 +68,17 @@ int kmeans(int myrank, const char *str, const point *pts, int n, int *res) {
     config.centroid_method = pt_centroid;
 
     /* Inputs for K-means */
-    config.objs = calloc(config.num_objs, sizeof(Pointer));
-    config.centers = calloc(config.k, sizeof(Pointer));
-    config.clusters = calloc(config.num_objs, sizeof(int));
+    config.objs = (Pointer*) calloc(config.num_objs, sizeof(Pointer));
+    config.centers = (Pointer*) calloc(config.k, sizeof(Pointer));
+    config.clusters = (int *) calloc(config.num_objs, sizeof(int));
 
     /* Storage for raw data */
 
-    init = calloc(config.k, sizeof(point));
+    init = (point *) calloc(config.k, sizeof(point));
 
     for (i = 0; i < n; i++) {
         /* Pointer to raw data */
-        config.objs[i] = &(pts[i]);
+        config.objs[i] = (Pointer) &(pts[i]);
     }
 
     /* Populate the initial means vector with random start points */
diff --git a/vendor/kmeans/kmeans_cl.c b/vendor/kmeans/kmeans_cl.c
@@ -274,7 +274,7 @@ kmeans_cl(kmeans_config *config)
 	 * Previous cluster state array. At this time, r doesn't mean anything
 	 * but it's ok
 	 */
-	clusters_last = kmeans_malloc(clusters_sz);
+	clusters_last = (int *) kmeans_malloc(clusters_sz);
 
 	while (1)
 	{
diff --git a/workflow/cwl/clt/clustering.cwl b/workflow/cwl/clt/clustering.cwl
@@ -23,6 +23,10 @@ inputs:
     inputBinding:
       position: 1
       prefix: -n
+  seed:
+    type: int
+    inputBinding:
+      position: 6
 outputs:
   indices:
     type: File
diff --git a/workflow/cwl/clt/silhouette.cwl b/workflow/cwl/clt/silhouette.cwl
@@ -0,0 +1,32 @@
+cwlVersion: v1.2
+class: CommandLineTool
+requirements:
+  ToolTimeLimit:
+    timelimit: 300
+arguments:
+  - position: 5
+    valueFrom: output.txt
+inputs:
+  annealing:
+    type: File
+    inputBinding:
+      position: 4
+  indices:
+    type: File
+    inputBinding:
+      position: 3
+  points:
+    type: File
+    inputBinding:
+      position: 2
+  silhouette:
+    type: File
+    inputBinding:
+      position: 1
+outputs:
+  output:
+    type: float
+    outputBinding:
+      glob: output.txt
+      loadContents: true
+      outputEval: $(parseFloat(self[0].contents))
diff --git a/workflow/cwl/config.yml b/workflow/cwl/config.yml
@@ -8,8 +8,12 @@ clustering_src:
     path: ../../Makefile
   - class: File
     path: ../../clustering.cpp
+  - class: File
+    path: ../../common.cpp
   - class: File
     path: ../../points.cpp
+  - class: File
+    path: ../../silhouette.cpp
   - class: Directory
     path: ../../include
   - class: Directory
diff --git a/workflow/cwl/main.cwl b/workflow/cwl/main.cwl
@@ -1,5 +1,9 @@
 cwlVersion: v1.2
 class: Workflow
+
+$namespaces:
+  cwltool: "http://commonwl.org/cwltool#"
+
 requirements:
   StepInputExpressionRequirement: {}
   ToolTimeLimit:
@@ -14,13 +18,15 @@ inputs:
       type: array
       items: [File, Directory]
   points: File
-  processes:
-    type: int
-    default: 3
+  processes: int?
+  threshold: float?
 outputs:
-  annealing:
+  output:
     type: File
-    outputSource: annealing/output
+    outputSource: loop/output
+  score:
+    type: float
+    outputSource: loop/score
 steps:
   build-clustering:
     run: clt/build.cwl
@@ -29,23 +35,81 @@ steps:
       output_path:
         valueFrom: build/bin/clustering
     out: [output]
-  clustering:
-    run: clt/clustering.cwl
-    in:
-      clustering: build-clustering/output
-      points: points
-      processes: processes
-    out: [indices, output]
   build-annealing:
     run: clt/build.cwl
     in:
       src: annealing_src
       output_path:
         valueFrom: build/bin/simAnnSingle.out
     out: [output]
-  annealing:
-    run: clt/annealing.cwl
+  build-silhouette:
+    run: clt/build.cwl
     in:
-      annealing: build-annealing/output
-      qubo: clustering/output
+      src: clustering_src
+      output_path:
+        valueFrom: build/bin/silhouette
     out: [output]
+  loop:
+    requirements:
+      InlineJavascriptRequirement: {}
+      cwltool:Loop:
+        loopWhen: $(inputs.score < inputs.threshold)
+        loop:
+          score: score
+          seed:
+            valueFrom: $(inputs.seed + 1)
+        outputMethod: last
+    run:
+      class: Workflow
+      inputs:
+        annealing_script: File
+        clustering_script: File
+        points: File
+        processes:
+          type: int
+          default: 3
+        seed: int
+        silhouette_script: File
+      outputs:
+        output:
+          type: File
+          outputSource: annealing/output
+        score:
+          type: float
+          outputSource: silhouette/output
+      steps:
+        clustering:
+          run: clt/clustering.cwl
+          in:
+            clustering: clustering_script
+            points: points
+            processes: processes
+            seed: seed
+          out: [indices, output]
+        annealing:
+          run: clt/annealing.cwl
+          in:
+            annealing: annealing_script
+            qubo: clustering/output
+          out: [output]
+        silhouette:
+          run: clt/silhouette.cwl
+          in:
+            silhouette: silhouette_script
+            points: points
+            indices: clustering/indices
+            annealing: annealing/output
+          out: [output]
+    in:
+      annealing_script: build-annealing/output
+      clustering_script: build-clustering/output
+      points: points
+      processes: processes
+      seed:
+        default: 0
+      score:
+        default: 0.0
+      silhouette_script: build-silhouette/output
+      threshold:
+        default: 0.5
+    out: [output, score]
diff --git a/workflow/streamflow.yml b/workflow/streamflow.yml
@@ -6,22 +6,30 @@ workflows:
       file: cwl/main.cwl
       settings: cwl/config.yml
     bindings:
+      - step: /build-annealing
+        target:
+          deployment: leonardo
+          service: booster
       - step: /build-clustering
         target:
           deployment: leonardo
           service: dcgp
-      - step: /clustering
+      - step: /build-silhouette
         target:
           deployment: leonardo
           service: dcgp
-      - step: /build-annealing
+      - step: /loop/clustering
         target:
           deployment: leonardo
-          service: booster
-      - step: /annealing
+          service: dcgp
+      - step: /loop/annealing
         target:
           deployment: leonardo
           service: booster
+      - step: /loop/silhouette
+        target:
+          deployment: leonardo
+          service: dcgp
 database:
   type: sqlite
   config:
@@ -39,7 +47,7 @@ deployments:
   leonardo:
     type: slurm
     config:
-      maxConcurrentJobs: 2
+      maxConcurrentJobs: 3
       services:
         booster:
           account: IscrC_SHPC-QC

Original file line number	Diff line number	Diff line change
`@@ -1,4 +1,5 @@`
`1`	`1`	`#include "common.h"`
	`2`	`+#include <algorithm>`
`2`	`3`	`#include <cmath>`
`3`	`4`	`#include <cstddef> // For offsetof`
`4`	`5`	`#include <fstream>`
Original file line number	Diff line number	Diff line change
`@@ -274,7 +274,7 @@ kmeans_cl(kmeans_config *config)`
`274`	`274`	`* Previous cluster state array. At this time, r doesn't mean anything`
`275`	`275`	`* but it's ok`
`276`	`276`	`*/`
`277`		`- clusters_last = kmeans_malloc(clusters_sz);`
	`277`	`+ clusters_last = (int *) kmeans_malloc(clusters_sz);`
`278`	`278`
`279`	`279`	`while (1)`
`280`	`280`	`{`