py2 test and cdl_sparql optimisation

Author: marqh

lib/bald/__init__.py

@@ -358,13 +358,14 @@ def unpack_predicate(self, astring):
                     result = astring.replace('{}__'.format(prefix),
                                              self.prefixes()[prefix])
         elif isinstance(astring, six.string_types):
-            predicate_alias_query = ('prefix dct: <http://purl.org/dc/terms/> '
-                                     'prefix owl: <http://www.w3.org/2002/07/owl#> '
-                                     'select ?uri where '
-                                  '{{?uri dct:identifier "{}" ; '
-                                  '      rdf:type ?type. '
-                                  'FILTER(?type in (rdf:Property, owl:ObjectProperty) ) '
-                                     '}}'.format(astring))
+            qstr = ('prefix dct: <http://purl.org/dc/terms/> \n'
+                    'prefix owl: <http://www.w3.org/2002/07/owl#> \n'
+                    'select ?uri where \n'
+                    '{{?uri dct:identifier "{}" ; \n'
+                    '      rdf:type ?type. \n'
+                    'FILTER(?type in (rdf:Property, owl:ObjectProperty) ) \n'
+                    '}}\n')
+            predicate_alias_query = (six.text_type(qstr).format(astring))
 
             qres = self.alias_graph.query(predicate_alias_query)
             results = list(qres)
@@ -389,12 +390,13 @@ def unpack_rdfobject(self, astring, predicate):
             #     msg = 'predicate must be a http uri, not {}'.format(predicate)
             #     raise ValueError(msg)
             # can be a file uri too
-            rdfobj_alias_query = ('prefix dct: <http://purl.org/dc/terms/> '
-                                  'select ?uri where '
-                                  '{{ <{pred}> rdfs:range ?range . '
-                                  '?uri dct:identifier "{id}" ; '
-                                  '     rdf:type ?range .'
-                                  '}}'.format(pred=predicate, id=astring))
+            qstr = ('prefix dct: <http://purl.org/dc/terms/> \n'
+                    'select ?uri where \n'
+                    '{{ <{pred}> rdfs:range ?range . \n'
+                    '?uri dct:identifier "{id}" ; \n'
+                    '     rdf:type ?range .\n'
+                    '}}\n')
+            rdfobj_alias_query = (six.text_type(qstr).format(pred=predicate, id=astring))
             # qres = self.alias_graph.query(rdfobj_alias_query)
             try:
                 qres = self.alias_graph.query(rdfobj_alias_query)

lib/bald/tests/integration/test_cdl_sparql.py

@@ -10,22 +10,29 @@
 from bald.tests import BaldTestCase
 from rdflib import Graph
 
+# a module level graph, to share for memory and performance
+thisGraph = [Graph()]
+loaded_boolean = []
+
 
 class Test(BaldTestCase):
     def setUp(self):
         self.cdl_path = os.path.join(os.path.dirname(__file__), 'CDL')
-        self.ttl_path = os.path.join(os.path.dirname(__file__), 'TTL')
-        self.graph = Graph()
 
-        #load bald graphs from cdl
-        for cdl_file in glob.glob(os.path.join(os.path.dirname(__file__), 'CDL', '*.cdl')):
-            with self.temp_filename('.nc') as tfile:
-               subprocess.check_call(['ncgen', '-o', tfile, cdl_file])
-               root_container = bald.load_netcdf(tfile, cache=self.acache)
-               curr_g = root_container.rdfgraph()
- 
-            #merge into graph in test obj
-            self.graph = self.graph  + curr_g
+        # Check to see if another test has already loaded the graph.
+        if not loaded_boolean:
+            # load bald graphs from cdl
+            for cdl_file in glob.glob(os.path.join(self.cdl_path, '*.cdl')):
+                with self.temp_filename('.nc') as tfile:
+                   subprocess.check_call(['ncgen', '-o', tfile, cdl_file])
+                   root_container = bald.load_netcdf(tfile, cache=self.acache)
+                   curr_g = root_container.rdfgraph()
+
+                #merge into graph in test obj
+                thisGraph[0] = thisGraph[0]  + curr_g
+            print('setting loaded_boolean')
+            loaded_boolean.append(True)
+        self.graph = thisGraph[0]
 
     def test_sparql_count_standard_names(self):
         #query standard_name values used and frequency