08.09 super() 函数

super() 函数

super(CurrentClassName, instance)

返回该类实例对应的父类对象。

 1class Leaf(object):
 2    def __init__(self, color="green"):
 3        self.color = color
 4    def fall(self):
 5        print "Splat!"
 6
 7class MapleLeaf(Leaf):
 8    def change_color(self):
 9        if self.color == "green":
10            self.color = "red"
11    def fall(self):
12        self.change_color()
13        super(MapleLeaf, self).fall()

这里,我们先改变树叶的颜色,然后再找到这个实例对应的父类,并调用父类的 fall() 方法:

1mleaf = MapleLeaf()
2
3print mleaf.color
4mleaf.fall()
5print mleaf.color
green
Splat!
red

回到我们的森林例子,这里我们将森林 Forest 作为父类,并定义一个子类 BurnableForest

 1import numpy as np
 2
 3class Forest(object):
 4    """ Forest can grow trees which eventually die."""
 5    def __init__(self, size=(150,150), p_sapling=0.0025):
 6        self.size = size
 7        self.trees = np.zeros(self.size, dtype=bool)
 8        self.p_sapling = p_sapling
 9        
10    def __repr__(self):
11        my_repr = "{}(size={})".format(self.__class__.__name__, self.size)
12        return my_repr
13    
14    def __str__(self):
15        return self.__class__.__name__
16    
17    @property
18    def num_cells(self):
19        """Number of cells available for growing trees"""
20        return np.prod(self.size)
21    
22    @property
23    def tree_fraction(self):
24        """
25        Fraction of trees
26        """
27        num_trees = self.trees.sum()
28        return float(num_trees) / self.num_cells
29    
30    def _rand_bool(self, p):
31        """
32        Random boolean distributed according to p, less than p will be True
33        """
34        return np.random.uniform(size=self.trees.shape) < p
35    
36    def grow_trees(self):
37        """
38        Growing trees.
39        """
40        growth_sites = self._rand_bool(self.p_sapling)
41        self.trees[growth_sites] = True    
42        
43    def advance_one_step(self):
44        """
45        Advance one step
46        """
47        self.grow_trees()
  • 将与燃烧相关的属性都被转移到了子类中去。
  • 修改两类的构造方法,将闪电概率放到子类的构造方法上,同时在子类的构造方法中,用 super 调用父类的构造方法。
  • 修改 advance_one_step(),父类中只进行生长,在子类中用 super 调用父类的 advance_one_step() 方法,并添加燃烧的部分。
 1class BurnableForest(Forest):
 2    """
 3    Burnable forest support fires
 4    """    
 5    def __init__(self, p_lightning=5.0e-6, **kwargs):
 6        super(BurnableForest, self).__init__(**kwargs)
 7        self.p_lightning = p_lightning        
 8        self.fires = np.zeros((self.size), dtype=bool)
 9    
10    def advance_one_step(self):
11        """
12        Advance one step
13        """
14        super(BurnableForest, self).advance_one_step()
15        self.start_fires()
16        self.burn_trees()
17        
18    @property
19    def fire_fraction(self):
20        """
21        Fraction of fires
22        """
23        num_fires = self.fires.sum()
24        return float(num_fires) / self.num_cells
25    
26    def start_fires(self):
27        """
28        Start of fire.
29        """
30        lightning_strikes = (self._rand_bool(self.p_lightning) & 
31            self.trees)
32        self.fires[lightning_strikes] = True
33        
34    def burn_trees(self):
35        """
36        Burn trees.
37        """
38        fires = np.zeros((self.size[0] + 2, self.size[1] + 2), dtype=bool)
39        fires[1:-1, 1:-1] = self.fires
40        north = fires[:-2, 1:-1]
41        south = fires[2:, 1:-1]
42        east = fires[1:-1, :-2]
43        west = fires[1:-1, 2:]
44        new_fires = (north | south | east | west) & self.trees
45        self.trees[self.fires] = False
46        self.fires = new_fires

测试父类:

1forest = Forest()
2
3forest.grow_trees()
4
5print forest.tree_fraction
0.00284444444444

测试子类:

1burnable_forest = BurnableForest()

调用自己和父类的方法:

1burnable_forest.grow_trees()
2burnable_forest.start_fires()
3burnable_forest.burn_trees()
4print burnable_forest.tree_fraction
0.00235555555556

查看变化:

 1import matplotlib.pyplot as plt
 2
 3%matplotlib inline
 4
 5forest = Forest()
 6forest2 = BurnableForest()
 7
 8tree_fractions = []
 9
10for i in range(2500):
11    forest.advance_one_step()
12    forest2.advance_one_step()
13    tree_fractions.append((forest.tree_fraction, forest2.tree_fraction))
14
15plt.plot(tree_fractions)
16
17plt.show()

png

__str____repr__self.__class__ 会根据类型不同而不同:

1forest
Forest(size=(150, 150))
1forest2
BurnableForest(size=(150, 150))
1print forest
Forest
1print forest2
BurnableForest