From 2cf62c5eba91866e7de828d9c5084aa3f21d4ba0 Mon Sep 17 00:00:00 2001
From: Lennart Regebro <regebro@gmail.com>
Date: Tue, 18 Sep 2018 09:12:19 +0200
Subject: [PATCH] See if we can get a better codeclimate score. (Probably not)

---
 xmldiff/diff.py       | 14 ++++++++------
 xmldiff/formatting.py | 10 ++++++++++
 xmldiff/main.py       | 22 ++++++++++++----------
 3 files changed, 30 insertions(+), 16 deletions(-)

diff --git a/xmldiff/diff.py b/xmldiff/diff.py
index 2585afa..cdef113 100644
--- a/xmldiff/diff.py
+++ b/xmldiff/diff.py
@@ -298,13 +298,15 @@ def find_pos(self, child):
                 break
         return i
 
+    def _get_matched_children(self, this, other, mapping):
+        return [child for child in this.getchildren()
+                if id(child) in mapping and
+                mapping[id(child)].getparent() is other]
+
     def align_children(self, left, right):
-        lchildren = [c for c in left.getchildren()
-                     if (id(c) in self._l2rmap and
-                         self._l2rmap[id(c)].getparent() is right)]
-        rchildren = [c for c in right.getchildren()
-                     if (id(c) in self._r2lmap and
-                         self._r2lmap[id(c)].getparent() is left)]
+        lchildren = self._get_matched_children(left, right, self._l2rmap)
+        rchildren = self._get_matched_children(right, left, self._r2lmap)
+
         if not lchildren or not rchildren:
             # Nothing to align
             return
diff --git a/xmldiff/formatting.py b/xmldiff/formatting.py
index 9752b20..a6f3e53 100644
--- a/xmldiff/formatting.py
+++ b/xmldiff/formatting.py
@@ -360,41 +360,51 @@ def _xpath(self, node, xpath):
         # one and exactly one element is found. This is to protect against
         # formatting a diff on the wrong tree, or against using ambigous
         # edit script xpaths.
+
+        # Check if this path is absolute (start with root)
         if xpath[0] == '/':
             root = True
             xpath = xpath[1:]
         else:
             root = False
 
+        # Find the first item in the path
         if '/' in xpath:
             path, rest = xpath.split('/', 1)
         else:
             path = xpath
             rest = ''
 
+        # Get the items index
         if '[' in path:
             path, index = path[:-1].split('[')
             index = int(index) - 1
             multiple = False
         else:
+            # If the path doesn't have an index, then [1] is implied
             index = 0
             multiple = True
 
         if root:
+            # Now add back the starting /, if there was one
             path = '/' + path
 
+        # Find the nodes that match
         matches = []
         for match in node.xpath(path, namespaces=node.nsmap):
             # Skip nodes that have been deleted
             if DELETE_NAME not in match.attrib:
                 matches.append(match)
 
+        # Get the node specified from the matches
         if index >= len(matches):
             raise ValueError('xpath %s[%s] not found at %s.' % (
                 path, index + 1, utils.getpath(node)))
         if len(matches) > 1 and multiple:
             raise ValueError('Multiple nodes found for xpath %s at %s.' % (
                 path, utils.getpath(node)))
+
+        # Found it!
         match = matches[index]
         if rest:
             return self._xpath(match, rest)
diff --git a/xmldiff/main.py b/xmldiff/main.py
index 4e290cf..530ced4 100644
--- a/xmldiff/main.py
+++ b/xmldiff/main.py
@@ -27,24 +27,26 @@ def diff_trees(left, right, F=None, uniqueattrs=None, formatter=None):
     return formatter.format(diffs, left)
 
 
-def diff_texts(left, right, F=None, uniqueattrs=None, formatter=None):
-    """Takes two Unicode strings containing XML"""
+def _parse_and_diff(tree_maker, left, right, F=None, uniqueattrs=None,
+                    formatter=None):
     normalize = bool(getattr(formatter, 'normalize', 1) & formatting.WS_TAGS)
     parser = etree.XMLParser(remove_blank_text=normalize)
-    left_tree = etree.fromstring(left, parser)
-    right_tree = etree.fromstring(right, parser)
+    left_tree = tree_maker(left, parser)
+    right_tree = tree_maker(right, parser)
     return diff_trees(left_tree, right_tree, F=F, uniqueattrs=uniqueattrs,
                       formatter=formatter)
 
 
+def diff_texts(left, right, F=None, uniqueattrs=None, formatter=None):
+    """Takes two Unicode strings containing XML"""
+    return _parse_and_diff(etree.fromstring, left, right, F,
+                           uniqueattrs, formatter)
+
+
 def diff_files(left, right, F=None, uniqueattrs=None, formatter=None):
     """Takes two filenames or streams, and diffs the XML in those files"""
-    normalize = bool(getattr(formatter, 'normalize', 1) & formatting.WS_TAGS)
-    parser = etree.XMLParser(remove_blank_text=normalize)
-    left_tree = etree.parse(left, parser)
-    right_tree = etree.parse(right, parser)
-    return diff_trees(left_tree, right_tree, F=F, uniqueattrs=uniqueattrs,
-                      formatter=formatter)
+    return _parse_and_diff(etree.parse, left, right, F,
+                           uniqueattrs, formatter)
 
 
 def make_parser():