refactor: candidate interpreation only on demand

Avoid creating interpretations not needed for the graph analysis task.

1 files changed, 3 insertions, 2 deletions

diff --git a/subprojects/language-web/src/main/java/tools/refinery/language/web/semantics/SemanticsService.java b/subprojects/language-web/src/main/java/tools/refinery/language/web/semantics/SemanticsService.java
index c828b3d5..2495430e 100644
--- a/subprojects/language-web/src/main/java/tools/refinery/language/web/semantics/SemanticsService.java
+++ b/subprojects/language-web/src/main/java/tools/refinery/language/web/semantics/SemanticsService.java
@@ -65,7 +65,8 @@ public class SemanticsService extends AbstractCachedService<SemanticsResult> {
                 var initializer = initializerProvider.get();
                 var builder = ModelStore.builder()
                                 .with(ViatraModelQueryAdapter.builder())
-                                .with(ReasoningAdapter.builder());
+                                .with(ReasoningAdapter.builder()
+                                                .requiredInterpretations(Concreteness.PARTIAL));
                 operationCanceledManager.checkCanceled(cancelIndicator);
                 try {
                         var modelSeed = initializer.createModel(problem, builder);
@@ -84,7 +85,7 @@ public class SemanticsService extends AbstractCachedService<SemanticsResult> {
                         }
                         return new SemanticsSuccessResult(nodeTrace, partialInterpretation);
                 } catch (RuntimeException e) {
-                        LOG.error("Error while computing semantics", e);
+                        LOG.debug("Error while computing semantics", e);
                         return new SemanticsErrorResult(e.getMessage());
                 }
         }