Merge 5a7cc0d into bef611b

CHANGES.rst

@@ -4,7 +4,7 @@ Changes
 2.5 (unreleased)
 ----------------
 
-- Nothing changed yet.
+- Make it possible to adjust the attributes considered when comparing nodes.
 
 
 2.4 (2019-10-09)

README.rst

@@ -84,5 +84,7 @@ Contributors
 
  * Albertas Agejevas, alga@shoobx.com
 
+ * Greg Kempe, greg@laws.africa
+
 The diff algorithm is based on "`Change Detection in Hierarchically Structured Information <http://ilpubs.stanford.edu/115/1/1995-46.pdf>`_",
 and the text diff is using Google's ``diff_match_patch`` algorithm.

tests/test_diff.py

@@ -1466,3 +1466,24 @@ def test_issue_21_default_namespaces(self):
         right = '<tag xmlns="ns">new</tag>'
         result = self._diff(left, right)
         self.assertEqual(result[0].node, "/*[1]")
+
+    def test_ignore_attribute(self):
+        # this differ ignores the attribute 'skip' when diffing
+        class IgnoringDiffer(Differ):
+            def node_attribs(self, node):
+                if 'skip' in node.attrib:
+                    attribs = dict(node.attrib)
+                    del attribs['skip']
+                    return attribs
+                return node.attrib
+
+        left = '<a><b foo="bar" skip="boom">text</b></a>'
+        right = '<a><b foo="bar" skip="different">text</b></a>'
+
+        parser = etree.XMLParser(remove_blank_text=True)
+        left_tree = etree.fromstring(left, parser)
+        right_tree = etree.fromstring(right, parser)
+        differ = IgnoringDiffer()
+        differ.set_trees(left_tree, right_tree)
+        editscript = list(differ.diff())
+        self.assertEqual(editscript, [])

xmldiff/diff.py

@@ -186,7 +186,7 @@ def node_text(self, node):
         texts = node.xpath("text()")
 
         # Then add attributes and values
-        for tag, value in sorted(node.attrib.items()):
+        for tag, value in sorted(self.node_attribs(node).items()):
             if tag[0] == "{":
                 tag = tag.split(
                     "}",
@@ -199,6 +199,10 @@ def node_text(self, node):
         self._text_cache[node] = result
         return result
 
+    def node_attribs(self, node):
+        """Return a dict of attributes to consider for this node."""
+        return node.attrib
+
     def leaf_ratio(self, left, right):
         # How similar two nodes are, with no consideration of their children
         # We use a simple ratio here, I tried Levenshtein distances
@@ -237,8 +241,8 @@ def update_node_attr(self, left, right):
 
         # Update: Look for differences in attributes
 
-        left_keys = set(left.attrib.keys())
-        right_keys = set(right.attrib.keys())
+        left_keys = set(self.node_attribs(left).keys())
+        right_keys = set(self.node_attribs(right).keys())
         new_keys = right_keys.difference(left_keys)
         removed_keys = left_keys.difference(right_keys)
         common_keys = left_keys.intersection(right_keys)