Fixed Swap Nodes in mutate_rows per issue #41 (#42)

* Fixed Swap Nodes in mutate_rows per issue #41

Swapped based on state, row[0], then sorted.  As per our conversation in issue #41 thread.

* Updated for fixed swap in mutation

Author: gaffney2010

docs/tutorial/index.rst

@@ -47,9 +47,9 @@ We can now evolve our population::
     Scoring Generation 2
     Generation 2 | Best Score: 2.1 0:C:0_C_0_C:0_D_1_C:1_C_1_D:1_D_1_D
     Scoring Generation 3
-    Generation 3 | Best Score: 2.2 0:C:0_C_0_C:0_D_1_D:1_C_1_D:1_D_1_D
+    Generation 3 | Best Score: 2.1 0:C:0_C_0_C:0_D_1_C:1_C_1_D:1_D_1_D
     Scoring Generation 4
-    Generation 4 | Best Score: 2.2 0:C:0_C_0_C:0_D_1_D:1_C_1_D:1_D_1_D
+    Generation 4 | Best Score: 2.1 0:C:0_C_0_C:0_D_1_C:1_C_1_D:1_D_1_D
 
 The :code:`run` command prints out the progress of the algorithm and this is
 also written to the output file (we passed :code:`output_filename` as an

src/axelrod_dojo/archetypes/fsm.py

@@ -74,10 +74,11 @@ def mutate_rows(rows, mutation_probability):
             n1 = randrange(nodes)
             n2 = randrange(nodes)
             for j, row in enumerate(rows):
-                if row[1] == n1:
-                    row[1] = n2
-                elif row[1] == n2:
-                    row[1] = n1
+                if row[0] == n1:
+                    row[0] = n2
+                elif row[0] == n2:
+                    row[0] = n1
+            rows.sort(key=lambda x: (x[0], 0 if x[1]==C else 1))
         return rows
 
     def mutate(self):