Fix highligher for trigrams; fix HTML escaping.

toolserver/copyvios/highlighter.py

@@ -6,6 +6,7 @@ from markupsafe import escape
 
 def highlight_delta(context, chain, delta):
     processed = []
+    dchain = delta.chain
     prev_prev = prev = chain.START
     i = 0
     all_words = chain.text.split()
@@ -15,14 +16,22 @@ def highlight_delta(context, chain, delta):
         words = paragraph.split(" ")
         for i, word in enumerate(words, i):
             try:
-                next = _strip_word(all_words[i+1])
+                next = _strip_word(all_words[i + 1])
+                try:
+                    next_next = _strip_word(all_words[i + 2])
+                except IndexError:
+                    next_next = chain.END
             except IndexError:
-                next = chain.END
+                next = next_next = chain.END
             sword = _strip_word(word)
-            block = (prev_prev, prev)  # Block for before
-            alock = (prev, sword)  # Block for after
-            before = [block in delta.chain and sword in delta.chain[block]]
-            after = [alock in delta.chain and next in delta.chain[alock]]
+            middle = (prev, sword) in dchain and next in dchain[(prev, sword)]
+            if middle:
+                before = after = True
+            else:
+                b_block = (prev_prev, prev)
+                a_block = (sword, next)
+                before = b_block in dchain and sword in dchain[b_block]
+                after = a_block in dchain and next_next in dchain[a_block]
             is_first = i == 0
             is_last = i + 1 == len(all_words)
             res = _highlight_word(word, before, after, is_first, is_last)
@@ -38,7 +47,7 @@ def _highlight_word(word, before, after, is_first, is_last):
         # Word is in the middle of a highlighted block, so don't change
         # anything unless this is the first word (force block to start) or the
         # last word (force block to end):
-        res = escape(word)
+        res = unicode(escape(word))
         if is_first:
             res = u'<span class="cv-hl">' + res
         if is_last:
@@ -59,18 +68,21 @@ def _highlight_word(word, before, after, is_first, is_last):
             res += u"</span>"
     else:
         # Word is completely outside of a highlighted block, so do nothing:
-        res = escape(word)
+        res = unicode(escape(word))
     return res
 
 def _fade_word(word, dir):
     if len(word) <= 4:
-        return u'<span class="cv-hl-{0}">{1}</span>'.format(dir, escape(word))
+        word = unicode(escape(word))
+        return u'<span class="cv-hl-{0}">{1}</span>'.format(dir, word)
     if dir == u"out":
+        before, after = unicode(escape(word[:-4])), unicode(escape(word[-4:]))
         base = u'{0}<span class="cv-hl-out">{1}</span>'
-        return base.format(escape(word[:-4]), escape(word[-4:]))
+        return base.format(before, after)
     else:
+        before, after = unicode(escape(word[:4])), unicode(escape(word[4:]))
         base = u'<span class="cv-hl-in">{0}</span>{1}'
-        return base.format(escape(word[:4]), escape(word[4:]))
+        return base.format(before, after)
 
 def _strip_word(word):
     return sub("[^\w\s-]", "", word.lower(), flags=UNICODE)