From e23a37dde87eb4b1db38256f60a748b8e73d4f98 Mon Sep 17 00:00:00 2001
From: Ben Kurtovic <ben.kurtovic@verizon.net>
Date: Tue, 4 Sep 2012 20:25:06 -0400
Subject: [PATCH] Join adjacent lines with only one word.

---
 toolserver/copyvios/highlighter.py | 18 ++++++++++++++++--
 1 file changed, 16 insertions(+), 2 deletions(-)

diff --git a/toolserver/copyvios/highlighter.py b/toolserver/copyvios/highlighter.py
index 1fa773c..539b3d7 100644
--- a/toolserver/copyvios/highlighter.py
+++ b/toolserver/copyvios/highlighter.py
@@ -22,9 +22,10 @@ def highlight_delta(context, chain, delta):
     i = degree
     numwords = len(chain.text.split())
     processed = []
-    for paragraph in chain.text.split("\n"):
+    paragraphs = chain.text.split("\n")
+    while paragraphs:
         words = []
-        for i, word in enumerate(paragraph.split(), i):
+        for i, word in enumerate(_get_next(paragraphs), i):
             if highlights[i]:
                 before = highlights[i - 1]
                 after = highlights[i + 1]
@@ -38,6 +39,19 @@ def highlight_delta(context, chain, delta):
 
     return u"<p>" + u"</p>\n<p>".join(processed) + u"</p>"
 
+def _get_next(paragraphs):
+    paragraph = paragraphs.pop(0)
+    body = paragraph.split()
+    if len(body) == 1:
+        while paragraphs:
+            next = paragraphs[0].split()
+            if len(next) == 1:
+                body += next
+                paragraphs.pop(0)
+            else:
+                break
+    return body
+
 def _highlight_word(word, before, after, first, last):
     if before and after:
         # Word is in the middle of a highlighted block: