From 64c440c198b4fa08c0909557c17b7003348f659d Mon Sep 17 00:00:00 2001
From: marwoodandrew <amarwood@aap.com.au>
Date: Mon, 20 Jun 2016 17:15:00 +1000
Subject: [PATCH] [SD-4728] Attempt to match the case of the original text in
 the Replace American Words macro

---
 server/aap/macros/replace_words.py      | 45 +++++++++++++++++++++++--
 server/aap/macros/replace_words_test.py | 18 +++++++---
 2 files changed, 56 insertions(+), 7 deletions(-)

diff --git a/server/aap/macros/replace_words.py b/server/aap/macros/replace_words.py
index b879bead8..7d5005f86 100644
--- a/server/aap/macros/replace_words.py
+++ b/server/aap/macros/replace_words.py
@@ -22,13 +22,52 @@ def find_and_replace(item, **kwargs):
     """
     diff = {}
 
+    def repl(new, old):
+        """
+        Returns a version of the "new" string that matches the case of the "old" string
+        :param new:
+        :param old:
+        :return: a string which is a version of "new" that matches the case of old.
+        """
+        if old.islower():
+            return new.lower()
+        elif old.isupper():
+            return new.upper()
+        else:
+            # the old string starts with upper case so we use the title function
+            if old[:1].isupper():
+                return new.title()
+            # it is more complex so try to match it
+            else:
+                result = ''
+                all_upper = True
+                for i, c in enumerate(old):
+                    if i >= len(new):
+                        break
+                    if c.isupper():
+                        result += new[i].upper()
+                    else:
+                        result += new[i].lower()
+                        all_upper = False
+                # append any remaining characters from new
+                if all_upper:
+                    result += new[i + 1:].upper()
+                else:
+                    result += new[i + 1:].lower()
+                return result
+
     def do_find_replace(input_string, words_list):
         for word in words_list:
             pattern = r'{}'.format(re.escape(word.get('existing', '')))
 
-            if re.search(pattern, input_string, flags=re.IGNORECASE):
-                diff[word.get('existing', '')] = word.get('replacement', '')
-                input_string = re.sub(pattern, word.get('replacement', ''), input_string, flags=re.IGNORECASE)
+            while re.search(pattern, input_string, flags=re.IGNORECASE):
+                # get the original string from the input
+                original = re.search(pattern, input_string, flags=re.IGNORECASE).group(0)
+                replacement = repl(word.get('replacement', ''), original)
+                if original == replacement:
+                    break
+                diff[original] = replacement
+                input_string = input_string.replace(original, replacement)
 
         return input_string
 
diff --git a/server/aap/macros/replace_words_test.py b/server/aap/macros/replace_words_test.py
index c7b16ca99..62e0f1d38 100644
--- a/server/aap/macros/replace_words_test.py
+++ b/server/aap/macros/replace_words_test.py
@@ -44,9 +44,9 @@ def test_find_replace_words_multiple_words(self):
 
         result, diff = self.under_test_find_and_replace(item)
         self.assertEqual(result['slugline'], 'this is colour')
-        self.assertEqual(result['headline'], 'colour is bad.')
+        self.assertEqual(result['headline'], 'Colour is bad.')
         self.assertEqual(result['body_html'], 'this is colour. tyre is great.')
-        self.assertDictEqual(diff, {'color': 'colour', 'tire': 'tyre'})
+        self.assertDictEqual(diff, {'ColOr': 'Colour', 'color': 'colour', 'tire': 'tyre'})
 
     def test_find_replace_words_same_words(self):
         item = {
@@ -54,8 +54,8 @@ def test_find_replace_words_same_words(self):
         }
 
         result, diff = self.under_test_find_and_replace(item)
-        self.assertEqual(result['body_html'], 'centre is great. centre is far')
-        self.assertDictEqual(diff, {'center': 'centre'})
+        self.assertEqual(result['body_html'], 'Centre is great. Centre is far')
+        self.assertDictEqual(diff, {'Center': 'Centre'})
 
     def test_find_replace_words(self):
         item = {
@@ -78,3 +78,13 @@ def test_find_replace_words_ending_with_dots(self):
         self.assertEqual(result['body_html'], 'George W Bush centre opens at 5 am')
         self.assertEqual(result['slugline'], 'George W Bush')
         self.assertDictEqual(diff, {'a.m.': 'am', 'George W. Bush': 'George W Bush', 'center': 'centre'})
+
+    def test_repetition_of_words_with_different_case(self):
+        item = {'body_html': 'Color color COLOR cOlOr'}
+        result, diff = self.under_test_find_and_replace(item)
+        self.assertEqual(result['body_html'], 'Colour colour COLOUR cOlOur')
+
+    def test_partial_word(self):
+        item = {'body_html': 'Something coloring it red'}
+        result, diff = self.under_test_find_and_replace(item)
+        self.assertEqual(result['body_html'], 'Something colouring it red')