I'm refactoring some code that a friend wrote and recently stumbled across this function:
def setup_parameters(self, data):
    '''Parse raw data to determine game settings.'''
    for line in data.split('\n'):
      line = line.strip().lower()
      if line:
        tokens = line.split()
        self.L.debug("tokens: " + str(tokens))
        key = tokens[0]
        if key == 'cols':
          self.width = int(tokens[1])
        elif key == 'rows':
          self.height = int(tokens[1])
        elif key == 'player_seed':
          random.seed(int(tokens[1]))
        elif key == 'turntime':
          self.turntime = int(tokens[1])
        elif key == 'loadtime':
          self.loadtime = int(tokens[1])
        elif key == 'viewradius2':
          self.viewradius2 = int(tokens[1])
        elif key == 'attackradius2':
          self.attackradius2 = int(tokens[1])
        elif key == 'spawnradius2':
          self.spawnradius2 = int(tokens[1])
As you can see, there is a nasty kind of switch statement here, that clearly calls for a dictionary. I'm tempted to write this as a class dictionary since the keys are constant, but since the keys map to attributes of an instance (ie, 'cols': self.width) this doesn't compile.
My question is then, what is the right way to refactor such code?
 
     
     
    