Move digest utilities to own class.

lakesuperior/connectors/graph_store_connector.py

@@ -75,41 +75,3 @@ class GraphStoreConnector:
         return self.ds.query(q, initBindings=initBindings, initNs=nsc)
 
 
-    ## PROTECTED METHODS ##
-
-    def _rdf_cksum(self, g):
-        '''
-        Generate a checksum for a graph.
-
-        This is not straightforward because a graph is derived from an
-        unordered data structure (RDF).
-
-        What this method does is ordering the graph by subject, predicate,
-        object, then creating a pickle string and a checksum of it.
-
-        N.B. The context of the triples is ignored, so isomorphic graphs would
-        have the same checksum regardless of the context(s) they are found in.
-
-        @TODO This can be later reworked to use a custom hashing algorithm.
-
-        @param rdflib.Graph g The graph to be hashed.
-
-        @return string SHA1 checksum.
-        '''
-        # Remove the messageDigest property, which at this point is very likely
-        # old.
-        g.remove((Variable('s'), nsc['premis'].messageDigest, Variable('o')))
-
-        ord_g = sorted(list(g), key=lambda x : (x[0], x[1], x[2]))
-        hash = sha1(pickle.dumps(ord_g)).hexdigest()
-
-        return hash
-
-
-    def _non_rdf_checksum(self, data):
-        '''
-        Generate a checksum of non-RDF content.
-
-        @TODO This can be later reworked to use a custom hashing algorithm.
-        '''
-        return sha1(data).hexdigest()

lakesuperior/util/digest.py

@@ -0,0 +1,43 @@
+class Digest:
+    '''
+    Various digest functions. May be merged into something more generic later.
+    '''
+    @staticmethod
+    def rdf_cksum(g):
+        '''
+        Generate a checksum for a graph.
+
+        This is not straightforward because a graph is derived from an
+        unordered data structure (RDF).
+
+        What this method does is ordering the graph by subject, predicate,
+        object, then creating a pickle string and a checksum of it.
+
+        N.B. The context of the triples is ignored, so isomorphic graphs would
+        have the same checksum regardless of the context(s) they are found in.
+
+        @TODO This can be later reworked to use a custom hashing algorithm.
+
+        @param rdflib.Graph g The graph to be hashed.
+
+        @return string SHA1 checksum.
+        '''
+        # Remove the messageDigest property, which at this point is very likely
+        # old.
+        g.remove((Variable('s'), nsc['premis'].messageDigest, Variable('o')))
+
+        ord_g = sorted(list(g), key=lambda x : (x[0], x[1], x[2]))
+        hash = sha1(pickle.dumps(ord_g)).hexdigest()
+
+        return hash
+
+
+    @staticmethod
+    def non_rdf_checksum(data):
+        '''
+        Generate a checksum of non-RDF content.
+
+        @TODO This can be later reworked to use a custom hashing algorithm.
+        '''
+        return sha1(data).hexdigest()
+