check non-serializable URIs in the sparql store

4 files changed, 16 insertions, 0 deletions

diff --git a/bsfs/triple_store/sparql/parse_filter.py b/bsfs/triple_store/sparql/parse_filter.py
index ff22de2..8959b2c 100644
--- a/bsfs/triple_store/sparql/parse_filter.py
+++ b/bsfs/triple_store/sparql/parse_filter.py
@@ -267,6 +267,8 @@ class Filter():
         """
         if not isinstance(node_type, bsc.Node):
             raise errors.BackendError(f'expected Node, found {node_type}')
+        if not rdflib.term._is_valid_uri(node.value): # pylint: disable=protected-access
+            raise errors.BackendError(f'<{node.value}> is not a serializable uri')
         return f'VALUES {head} {{ <{node.value}> }}'
 
     def _equals(self, node_type: bsc.Vertex, node: ast.filter.Equals, head: str) -> str:
diff --git a/bsfs/triple_store/sparql/sparql.py b/bsfs/triple_store/sparql/sparql.py
index 5890bcc..bd98f46 100644
--- a/bsfs/triple_store/sparql/sparql.py
+++ b/bsfs/triple_store/sparql/sparql.py
@@ -284,6 +284,9 @@ class SparqlStore(base.TripleStoreBase):
             raise errors.ConsistencyError(f'{node_type} is not defined in the schema')
         # check and create guids
         for guid in guids:
+            # check convert to rdflib.URIRef
+            if not rdflib.term._is_valid_uri(guid): # pylint: disable=protected-access
+                raise ValueError(guids)
             subject = rdflib.URIRef(guid)
             # check node existence
             if (subject, rdflib.RDF.type, None) in self._graph:
@@ -324,6 +327,9 @@ class SparqlStore(base.TripleStoreBase):
         # check guids
         # FIXME: Fail or skip inexistent nodes?
         guids = set(guids)
+        invalid = {guid for guid in guids if not rdflib.term._is_valid_uri(guid)} # pylint: disable=protected-access
+        if len(invalid) > 0:
+            raise ValueError(invalid)
         inconsistent = {guid for guid in guids if not self._has_type(guid, node_type)}
         if len(inconsistent) > 0:
             raise errors.InstanceError(inconsistent)
diff --git a/test/triple_store/sparql/test_parse_filter.py b/test/triple_store/sparql/test_parse_filter.py
index 8a9940e..6db9224 100644
--- a/test/triple_store/sparql/test_parse_filter.py
+++ b/test/triple_store/sparql/test_parse_filter.py
@@ -157,6 +157,8 @@ class TestParseFilter(unittest.TestCase):
     def test_is(self):
         # _is requires a node
         self.assertRaises(errors.BackendError, self.parser._is, self.schema.literal(ns.bsfs.Literal), ast.filter.Is('http://example.com/entity#1234'), '?ent')
+        # _is requires a serializable guid
+        self.assertRaises(errors.BackendError, self.parser._is, self.schema.node(ns.bsfs.Entity), ast.filter.Is('http://example.com/entity#foo and bar'), '?ent')
         # a single Is statement
         q = self.parser(self.schema.node(ns.bsfs.Entity), ast.filter.Is('http://example.com/entity#1234'))
         self.assertSetEqual({str(guid) for guid, in q(self.graph)},
diff --git a/test/triple_store/sparql/test_sparql.py b/test/triple_store/sparql/test_sparql.py
index b1d99ac..d880082 100644
--- a/test/triple_store/sparql/test_sparql.py
+++ b/test/triple_store/sparql/test_sparql.py
@@ -678,6 +678,9 @@ class TestSparqlStore(unittest.TestCase):
         self.assertRaises(errors.ConsistencyError, store.create, self.schema.node(ns.bsfs.Entity).child(ns.bsfs.invalid), {
             URI('http://example.com/me/entity#1234'), URI('http://example.com/me/entity#4321')})
 
+        # guid must be valid
+        self.assertRaises(ValueError, store.create, self.schema.node(ns.bsfs.Entity), {URI('http://example.com/me/foo and bar')})
+
         # can create some nodes
         ent_type = store.schema.node(ns.bsfs.Entity)
         store.create(ent_type, {URI('http://example.com/me/entity#1234'), URI('http://example.com/me/entity#4321')})
@@ -766,6 +769,9 @@ class TestSparqlStore(unittest.TestCase):
         # invalid predicate is not permitted
         self.assertRaises(errors.ConsistencyError, store.set, ent_type, ent_ids, p_invalid, {'http://example.com/me/tag#1234'})
 
+        # invalid guid is not permitted
+        self.assertRaises(ValueError, store.set, ent_type, {URI('http://example.com/me/foo and bar')}, p_filesize, {1234})
+
         # predicate must match node_type
         self.assertRaises(errors.ConsistencyError, store.set, tag_type, tag_ids, p_filesize, {1234})