Optimize tween

* Move the floating point division out of the inner loop
* Use map() / list comprehensions where possible, in preference to nested loop
* Don't call show() explicitly if _set() already did it
* Regression test (not committed) shows identical _values being written at each step

library/piglow.py

@@ -1,4 +1,7 @@
+from __future__ import division
+
 import atexit
+import operator
 import time
 from sys import exit
 
@@ -250,16 +253,17 @@ def tween(duration, end, start=None):
     if start is None:
         start = _values.copy()
 
+    deltas = map(operator.sub, end, start)
+    deltas_per_frame = [d / steps for d in deltas]
+
+    new = start
     for x in range(steps):
-        new = []
-        for y in range(18):
-            s = start[y]
-            e = end[y]
-            c = float(e - s)
-            b = s + ((c / float(steps)) * (x + 1))
-            new.append(int(b))
-        _set(0, new)
-        show()
+        new = list(map(operator.add, new, deltas_per_frame))
+        new_ints = [round(n) for n in new]
+        _set(0, new_ints)
+        # avoid double write
+        if not auto_update:
+            show()
         time.sleep(fps)