@MirrorIce your tests are wrong; they were wrong even in the previous build and should have actually failed; You are missing a hypotheses generator from (almost) each of the pipelines. You should run PointCloudClusterExtractor so annotators like Primitive shape, feature extractor etc. have some clusters to annotate; (explains the low percentage on the coverage as well;;;basically the algorithms are never really executed since there are no clusters)
@MirrorIce your tests are wrong; they were wrong even in the previous build and should have actually failed; You are missing a hypotheses generator from (almost) each of the pipelines. You should run PointCloudClusterExtractor so annotators like Primitive shape, feature extractor etc. have some clusters to annotate; (explains the low percentage on the coverage as well;;;basically the algorithms are never really executed since there are no clusters)