test for directories

Author: stain

cwltool/provenance.py

@@ -636,7 +636,9 @@ def declare_artefact(self, value):
 
                 # FIXME: .listing might not be populated yet - hopefully
                 # a later call to this method will sort that
+                is_empty = True
                 for f in value.get("listing", []):
+                    is_empty = False
                     # Declare child-artifacts
                     entity = self.declare_artefact(f)
                     self.document.membership(coll, entity)
@@ -686,7 +688,7 @@ def declare_artefact(self, value):
                     ore_doc.serialize(provenance_file, format="rdf", rdf_format="turtle")
                 self.research_object.add_annotation(dir_id, [ore_doc_fn], ORE["isDescribedBy"].uri)
 
-                if not coll_attribs:
+                if is_empty:
                     # Empty directory
                     coll.add_asserted_type(PROV["EmptyCollection"])
                     coll.add_asserted_type(PROV["EmptyDictionary"])

tests/test_provenance.py

@@ -361,6 +361,37 @@ def check_prov(self, nested=False, single_tool=False, directory=False):
                 directories = set(g.subjects(RDF.type, RO.Folder))
                 self.assertTrue(directories)
 
+                for d in directories:
+                    self.assertTrue((d,RDF.type,PROV.Dictionary) in g)
+                    self.assertTrue((d,RDF.type,PROV.Collection) in g)
+                    self.assertTrue((d,RDF.type,PROV.Entity) in g)
+
+                    files = set()
+                    for entry in g.objects(d, PROV.hadDictionaryMember):
+                        self.assertTrue((entry,RDF.type,PROV.KeyEntityPair) in g)
+                        # We don't check what that filename is here
+                        self.assertTrue(set(g.objects(entry,PROV.pairKey)))
+
+                        # RO:Folder aspect
+                        self.assertTrue(set(g.objects(entry,RO.entryName)))
+                        self.assertTrue((d,ORE.aggregates,entry) in g)
+                        self.assertTrue((entry,RDF.type,RO.FolderEntry) in g)
+                        self.assertTrue((entry,RDF.type,ORE.Proxy) in g)
+                        self.assertTrue((entry,ORE.proxyIn,d) in g)
+                        self.assertTrue((entry,ORE.proxyIn,d) in g)
+
+                        # Which file?
+                        entities = set(g.objects(entry, PROV.pairEntity))
+                        self.assertTrue(entities)
+                        f = entities.pop()
+                        files.add(f)
+                        self.assertTrue((entry,ORE.proxyFor,f) in g)
+                        self.assertTrue((f,RDF.type,PROV.Entity) in g)
+
+                    if not files:
+                        self.assertTrue((d,RDF.type,PROV.EmptyCollection) in g)
+                        self.assertTrue((d,RDF.type,PROV.EmptyDictionary) in g)
+
 
 class TestConvertPath(unittest.TestCase):
     def test_nt_to_posix(self):