Merge lp:~arthur-darcet/beautifulsoup/beautifulsoup into lp:beautifulsoup

[Arthur Darcet]

Hi,

We use BeautifulSoup in a program that calls PageElement.unwrap and PageElement.replace_with a lot, and those two functions trigger two calls to `self.parent.index(self)` where only one would be needed.
I amended `PageElement.extract` to take an optional (and internal) parameter if the tag index in its parent is already known. For our use-case, all the processing time is spent in `list.index`, so this patch is a 2x speed up.

Let me know if this isn't the right place to submit a patch

Arthur Darcet

[petrescu (paul-bogdan)]

Needs reviewer' ack,

Paul

=== modified file 'bs4/element.py'
--- bs4/element.py	2016-07-26 17:03:59 +0000
+++ bs4/element.py	2017-04-26 13:36:26 +0000
@@ -240,7 +240,7 @@
             raise ValueError("Cannot replace a Tag with its parent.")
         old_parent = self.parent
         my_index = self.parent.index(self)
-        self.extract()
+        self.extract(_self_index=my_index)
         old_parent.insert(my_index, replace_with)
         return self
     replaceWith = replace_with  # BS3
@@ -252,7 +252,7 @@
                 "Cannot replace an element with its contents when that"
                 "element is not part of a tree.")
         my_index = self.parent.index(self)
-        self.extract()
+        self.extract(_self_index=my_index)
         for child in reversed(self.contents[:]):
             my_parent.insert(my_index, child)
         return self
@@ -264,10 +264,12 @@
         wrap_inside.append(me)
         return wrap_inside
 
-    def extract(self):
+    def extract(self, _self_index=None):
         """Destructively rips this element out of the tree."""
         if self.parent is not None:
-            del self.parent.contents[self.parent.index(self)]
+            if _self_index is None:
+                _self_index = self.parent.index(self)
+            del self.parent.contents[_self_index]
 
         #Find the two elements that would be next to each other if
         #this element (and any children) hadn't been parsed. Connect