Permuting Sequences

The functions defined in Example 17-23 shuffle sequences in a number of ways:

  • permute constructs a list with all valid permutations of any sequence.
  • subset constructs a list with all valid permutations of a specific length.
  • combo works like subset, but order doesn't matter: permutations of the same items are filtered out.

These results are useful in a variety of algorithms: searches, statistical analysis, and more. For instance, one way to find an optimal ordering for items is to put them in a list, generate all possible permutations, and simply test each one in turn. All three of the functions make use of the generic sequence slicing tricks of the reversal functions in the prior section, so that the result list contains sequences of the same type as the one passed in (e.g., when we permute a string, we get back a list of strings).

Example 17-23. PP2EDstructClassicspermcomb.py

def permute(list):
 if not list: # shuffle any sequence
 return [list] # empty sequence
 else:
 res = []
 for i in range(len(list)):
 rest = list[:i] + list[i+1:] # delete current node
 for x in permute(rest): # permute the others
 res.append(list[i:i+1] + x) # add node at front
 return res

def subset(list, size):
 if size == 0 or not list: # order matters here
 return [list[:0]] # an empty sequence
 else:
 result = []
 for i in range(len(list)):
 pick = list[i:i+1] # sequence slice
 rest = list[:i] + list[i+1:] # keep [:i] part
 for x in subset(rest, size-1):
 result.append(pick + x)
 return result

def combo(list, size):
 if size == 0 or not list: # order doesn't matter
 return [list[:0]] # xyz == yzx
 else:
 result = []
 for i in range(0, (len(list) - size) + 1): # iff enough left
 pick = list[i:i+1] 
 rest = list[i+1:] # drop [:i] part
 for x in combo(rest, size - 1):
 result.append(pick + x)
 return result

As in the reversal functions, all three of these work on any sequence object that supports len, slicing, and concatenation operations. For instance, we can use permute on instances of some of the stack classes defined at the start of this chapter; we'll get back a list of stack instance objects with shuffled nodes.

Here are our sequence shufflers in action. Permuting a list enables us to find all the ways the items can be arranged. For instance, for a four-item list, there are 24 possible permutations (4 x 3 x 2 x 1). After picking one of the four for the first position, there are only three left to choose from for the second, and so on. Order matters: [1,2,3] is not the same as [1,3,2], so both appear in the result:

C:...PP2EDstructClassics>python
>>> from permcomb import *
>>> permute([1,2,3])
[[1, 2, 3], [1, 3, 2], [2, 1, 3], [2, 3, 1], [3, 1, 2], [3, 2, 1]]
>>> permute('abc')
['abc', 'acb', 'bac', 'bca', 'cab', 'cba']
>>> permute('help')
['help', 'hepl', 'hlep', 'hlpe', 'hpel', 'hple', 'ehlp', 'ehpl', 'elhp', 'elph',
 'ephl', 'eplh', 'lhep', 'lhpe', 'lehp', 'leph', 'lphe', 'lpeh', 'phel', 'phle',
 'pehl', 'pelh', 'plhe', 'pleh']

combo results are related to permutations, but a fixed-length constraint is put on the result, and order doesn't matter: abc is the same as acb, so only one is added to the result set:

>>> combo([1,2,3], 3)
[[1, 2, 3]]
>>> combo('abc', 3)
['abc']
>>> combo('abc', 2)
['ab', 'ac', 'bc']
>>> combo('abc', 4)
[]
>>> combo((1, 2, 3, 4), 3)
[(1, 2, 3), (1, 2, 4), (1, 3, 4), (2, 3, 4)]
>>> for i in range(0, 6): print i, combo("help", i)
... 
0 ['']
1 ['h', 'e', 'l', 'p']
2 ['he', 'hl', 'hp', 'el', 'ep', 'lp']
3 ['hel', 'hep', 'hlp', 'elp']
4 ['help']
5 []

Finally, subset is just fixed-length permutations; order matters, so the result is larger than for combo. In fact, calling subset with the length of the sequence is identical to permute:

>>> subset([1,2,3], 3)
[[1, 2, 3], [1, 3, 2], [2, 1, 3], [2, 3, 1], [3, 1, 2], [3, 2, 1]]
>>> subset('abc', 3)
['abc', 'acb', 'bac', 'bca', 'cab', 'cba']
>>> for i in range(0, 6): print i, subset("help", i)
... 
0 ['']
1 ['h', 'e', 'l', 'p']
2 ['he', 'hl', 'hp', 'eh', 'el', 'ep', 'lh', 'le', 'lp', 'ph', 'pe', 'pl']
3 ['hel', 'hep', 'hle', 'hlp', 'hpe', 'hpl', 'ehl', 'ehp', 'elh', 'elp', 'eph',
 'epl', 'lhe', 'lhp', 'leh', 'lep', 'lph', 'lpe', 'phe', 'phl', 'peh', 'pel',
 'plh', 'ple']
4 ['help', 'hepl', 'hlep', 'hlpe', 'hpel', 'hple', 'ehlp', 'ehpl', 'elhp', 
 'elph', 'ephl', 'eplh', 'lhep', 'lhpe', 'lehp', 'leph', 'lphe', 'lpeh', 
 'phel', 'phle', 'pehl', 'pelh', 'plhe', 'pleh']
5 ['help', 'hepl', 'hlep', 'hlpe', 'hpel', 'hple', 'ehlp', 'ehpl', 'elhp', 
 'elph', 'ephl', 'eplh', 'lhep', 'lhpe', 'lehp', 'leph', 'lphe', 'lpeh', 
 'phel', 'phle', 'pehl', 'pelh', 'plhe', 'pleh']

Introducing Python

Part I: System Interfaces

System Tools

Parallel System Tools

Larger System Examples I

Larger System Examples II

Part II: GUI Programming

Graphical User Interfaces

A Tkinter Tour, Part 1

A Tkinter Tour, Part 2

Larger GUI Examples

Part III: Internet Scripting

Network Scripting

Client-Side Scripting

Server-Side Scripting

Larger Web Site Examples I

Larger Web Site Examples II

Advanced Internet Topics

Part IV: Assorted Topics

Databases and Persistence

Data Structures

Text and Language

Part V: Integration

Extending Python

Embedding Python

VI: The End

Conclusion Python and the Development Cycle



Programming Python
Python Programming for the Absolute Beginner, 3rd Edition
ISBN: 1435455002
EAN: 2147483647
Year: 2000
Pages: 245

Flylib.com © 2008-2020.
If you may any questions please contact us: flylib@qtcs.net