Mostly in SequenceMatcher.{__chain_b, find_longest_match}:

This now does a dynamic analysis of which elements are so frequently
repeated as to constitute noise.  The primary benefit is an enormous
speedup in find_longest_match, as the innermost loop can have factors
of 100s less potential matches to worry about, in cases where the
sequences have many duplicate elements.  In effect, this zooms in on
sequences of non-ubiquitous elements now.

While I like what I've seen of the effects so far, I still consider
this experimental.  Please give it a try!

1 files changed, 55 insertions, 18 deletions

diff --git a/Lib/difflib.py b/Lib/difflib.py
index 69002f9..8fd2561 100644
--- a/Lib/difflib.py
+++ b/Lib/difflib.py
@@ -161,8 +161,6 @@ class SequenceMatcher:
         # b2j
         #      for x in b, b2j[x] is a list of the indices (into b)
         #      at which x appears; junk elements do not appear
-        # b2jhas
-        #      b2j.has_key
         # fullbcount
         #      for x in b, fullbcount[x] == the number of times x
         #      appears in b; only materialized if really needed (used
@@ -188,6 +186,10 @@ class SequenceMatcher:
         #      for x in b, isbjunk(x) == isjunk(x) but much faster;
         #      it's really the has_key method of a hidden dict.
         #      DOES NOT WORK for x in a!
+        # isbpopular
+        #      for x in b, isbpopular(x) is true iff b is reasonably long
+        #      (at least 200 elements) and x accounts for more than 1% of
+        #      its elements.  DOES NOT WORK for x in a!
 
         self.isjunk = isjunk
         self.a = self.b = None
@@ -266,6 +268,12 @@ class SequenceMatcher:
     # map at all, which stops the central find_longest_match method
     # from starting any matching block at a junk element ...
     # also creates the fast isbjunk function ...
+    # b2j also does not contain entries for "popular" elements, meaning
+    # elements that account for more than 1% of the total elements, and
+    # when the sequence is reasonably large (>= 200 elements); this can
+    # be viewed as an adaptive notion of semi-junk, and yields an enormous
+    # speedup when, e.g., comparing program files with hundreds of
+    # instances of "return NULL;" ...
     # note that this is only called when b changes; so for cross-product
     # kinds of matches, it's best to call set_seq2 once, then set_seq1
     # repeatedly
@@ -282,25 +290,36 @@ class SequenceMatcher:
         # out the junk later is much cheaper than building b2j "right"
         # from the start.
         b = self.b
+        n = len(b)
         self.b2j = b2j = {}
-        self.b2jhas = b2jhas = b2j.has_key
-        for i in xrange(len(b)):
-            elt = b[i]
-            if b2jhas(elt):
-                b2j[elt].append(i)
+        populardict = {}
+        for i, elt in enumerate(b):
+            if elt in b2j:
+                indices = b2j[elt]
+                if n >= 200 and len(indices) * 100 > n:
+                    populardict[elt] = 1
+                    del indices[:]
+                else:
+                    indices.append(i)
             else:
                 b2j[elt] = [i]
 
+        # Purge leftover indices for popular elements.
+        for elt in populardict:
+            del b2j[elt]
+
         # Now b2j.keys() contains elements uniquely, and especially when
         # the sequence is a string, that's usually a good deal smaller
         # than len(string).  The difference is the number of isjunk calls
         # saved.
-        isjunk, junkdict = self.isjunk, {}
+        isjunk = self.isjunk
+        junkdict = {}
         if isjunk:
-            for elt in b2j.keys():
-                if isjunk(elt):
-                    junkdict[elt] = 1   # value irrelevant; it's a set
-                    del b2j[elt]
+            for d in populardict, b2j:
+                for elt in d.keys():
+                    if isjunk(elt):
+                        junkdict[elt] = 1
+                        del d[elt]
 
         # Now for x in b, isjunk(x) == junkdict.has_key(x), but the
         # latter is much faster.  Note too that while there may be a
@@ -308,6 +327,7 @@ class SequenceMatcher:
         # elements is probably small.  So the memory burden of keeping
         # this dict alive is likely trivial compared to the size of b2j.
         self.isbjunk = junkdict.has_key
+        self.isbpopular = populardict.has_key
 
     def find_longest_match(self, alo, ahi, blo, bhi):
         """Find longest matching block in a[alo:ahi] and b[blo:bhi].
@@ -388,6 +408,19 @@ class SequenceMatcher:
                     besti, bestj, bestsize = i-k+1, j-k+1, k
             j2len = newj2len
 
+        # Extend the best by non-junk elements on each end.  In particular,
+        # "popular" non-junk elements aren't in b2j, which greatly speeds
+        # the inner loop above, but also means "the best" match so far
+        # doesn't contain any junk *or* popular non-junk elements.
+        while besti > alo and bestj > blo and \
+              not isbjunk(b[bestj-1]) and \
+              a[besti-1] == b[bestj-1]:
+            besti, bestj, bestsize = besti-1, bestj-1, bestsize+1
+        while besti+bestsize < ahi and bestj+bestsize < bhi and \
+              not isbjunk(b[bestj+bestsize]) and \
+              a[besti+bestsize] == b[bestj+bestsize]:
+            bestsize += 1
+
         # Now that we have a wholly interesting match (albeit possibly
         # empty!), we may as well suck up the matching junk on each
         # side of it too.  Can't think of a good reason not to, and it
@@ -736,12 +769,16 @@ class Differ:
         - `linejunk`: A function that should accept a single string argument,
           and return true iff the string is junk. The module-level function
           `IS_LINE_JUNK` may be used to filter out lines without visible
-          characters, except for at most one splat ('#').
+          characters, except for at most one splat ('#').  It is recommended
+          to leave linejunk None; as of Python 2.3, the underlying
+          SequenceMatcher class has grown an adaptive notion of "noise" lines
+          that's better than any static definition the author has ever been
+          able to craft.
 
         - `charjunk`: A function that should accept a string of length 1. The
           module-level function `IS_CHARACTER_JUNK` may be used to filter out
           whitespace characters (a blank or tab; **note**: bad idea to include
-          newline in this!).
+          newline in this!).  Use of IS_CHARACTER_JUNK is recommended.
         """
 
         self.linejunk = linejunk
@@ -1005,7 +1042,7 @@ def IS_CHARACTER_JUNK(ch, ws=" \t"):
 
 del re
 
-def ndiff(a, b, linejunk=IS_LINE_JUNK, charjunk=IS_CHARACTER_JUNK):
+def ndiff(a, b, linejunk=None, charjunk=IS_CHARACTER_JUNK):
     r"""
     Compare `a` and `b` (lists of strings); return a `Differ`-style delta.
 
@@ -1013,9 +1050,9 @@ def ndiff(a, b, linejunk=IS_LINE_JUNK, charjunk=IS_CHARACTER_JUNK):
     functions (or None):
 
     - linejunk: A function that should accept a single string argument, and
-      return true iff the string is junk. The default is module-level function
-      IS_LINE_JUNK, which filters out lines without visible characters, except
-      for at most one splat ('#').
+      return true iff the string is junk.  The default is None, and is
+      recommended; as of Python 2.3, an adaptive notion of "noise" lines is
+      used that does a good job on its own.
 
     - charjunk: A function that should accept a string of length 1. The
       default is module-level function IS_CHARACTER_JUNK, which filters out