3.4.3 Traits是什么

trait是可以用于常规Python对象属性的类型定义，给出属性的一些额外特性:

  • 标准化:
    • 初始化
    • 验证
    • 推迟
  • 通知
  • 可视化
  • 文档

类可以自由混合基于trait的属性与通用Python属性，或者选择允许在这个类中只使用固定的或开放的trait属性集。类定义的Trait属性自动继承自由这个类衍生的其他子类。

创建一个traits类的常用方式是通过扩展HasTraits基础类，并且定义类的traits :

In [1]:

  1. from traits.api import HasTraits, Str, Float
  3. class Reservoir(HasTraits):
  5.     name = Str
  6.     max_storage = Float

对Traits 3.x用户来说

如果使用Traits 3.x, 你需要调整traits包的命名空间:

  • traits.api应该为enthought.traits.api
  • traitsui.api应该为enthought.traits.ui.api

像这样使用traits类和使用其他Python类一样简单。注意，trait值通过关键词参数传递:

In [2]:

  1. reservoir = Reservoir(name='Lac de Vouglans', max_storage=605)

3.4.3.1 初始化

所有的traits都有一个默认值来初始化变量。例如，基础python类型有如下的trait等价物:

TraitPython类型内置默认值
BoolBooleanFalse
ComplexComplex number0+0j
FloatFloating point number0.0
IntPlain integer0
LongLong integer0L
StrString''
UnicodeUnicodeu''

存在很多其他预定义的trait类型: Array, Enum, Range, Event, Dict, List, Color, Set, Expression, Code, Callable, Type, Tuple, etc。

自定义默认值可以在代码中定义:

In [3]:

  1. from traits.api import HasTraits, Str, Float
  3. class Reservoir(HasTraits):
  5.     name = Str
  6.     max_storage = Float(100)
  8. reservoir = Reservoir(name='Lac de Vouglans')

复杂初始化

当一个trait需要复杂的初始化时，可以实施XXX默认魔法方法。当调用XXX trait时，它会被懒惰的调用。例如：

In [4]:

  1. def _name_default(self):
  2.     """ Complex initialisation of the reservoir name. """
  4.     return 'Undefined'

3.4.3.2 验证

当用户试图设置trait的内容时，每一个trait都会被验证:

In [5]:

  1. reservoir = Reservoir(name='Lac de Vouglans', max_storage=605)
  3. reservoir.max_storage = '230'
  1.  ------------- ------------- ------------- ------------- -----------------------
  2. TraitError                                Traceback (most recent call last)
  3. <ipython-input-5-cbed071af0b9> in <module>()
  4.  1 reservoir = Reservoir(name='Lac de Vouglans', max_storage=605)
  5.  2 
  6. ----> 3  reservoir.max_storage = '230'
  8. /Library/Python/2.7/site-packages/traits/trait_handlers.pyc in error(self, object, name, value)
  9.  170         """
  10.  171         raise TraitError( object, name, self.full_info( object, name, value ),
  11. --> 172 value ) 173 
  12.  174     def full_info ( self, object, name, value ):
  14. TraitError: The 'max_storage' trait of a Reservoir instance must be a float, but a value of '230' <type 'str'> was specified.

3.4.3.3 文档

从本质上说，所有的traits都提供关于模型自身的文档。创建类的声明方式使它是自解释的:

In [6]:

  1. from traits.api import HasTraits, Str, Float
  3. class Reservoir(HasTraits):
  5.     name = Str
  6.     max_storage = Float(100)

trait的desc元数据可以用来提供关于trait更多的描述信息:

In [7]:

  1. from traits.api import HasTraits, Str, Float
  3. class Reservoir(HasTraits):
  5.     name = Str
  6.     max_storage = Float(100, desc='Maximal storage [hm3]')

现在让我们来定义完整的reservoir类:

In [8]:

  1. from traits.api import HasTraits, Str, Float, Range
  3. class Reservoir(HasTraits):
  4.     name = Str
  5.     max_storage = Float(1e6, desc='Maximal storage [hm3]')
  6.     max_release = Float(10, desc='Maximal release [m3/s]')
  7.     head = Float(10, desc='Hydraulic head [m]')
  8.     efficiency = Range(0, 1.)
  10.     def energy_production(self, release):
  11.         ''' Returns the energy production [Wh] for the given release [m3/s]
  12.  '''
  13.         power = 1000 * 9.81 * self.head * release * self.efficiency
  14.         return power * 3600
  16. if __name__ == '__main__':
  17.     reservoir = Reservoir(
  18.                         name = 'Project A',
  19.                         max_storage = 30,
  20.                         max_release = 100.0,
  21.                         head = 60,
  22.                         efficiency = 0.8
  23.                     )
  25.     release = 80
  26.     print 'Releasing {} m3/s produces {} kWh'.format(
  27.                         release, reservoir.energy_production(release)
  28.                     )
  1. Releasing 80 m3/s produces 1.3561344e+11 kWh

3.4.3.4 可视化: 打开一个对话框

Traits库也关注用户界面，可以弹出一个Reservoir类的默认视图:

In [ ]:

  1. reservoir1 = Reservoir()
  2. reservoir1.edit_traits()

3.4.3 Traits是什么 - 图1

TraitsUI简化了创建用户界面的方式。HasTraits类上的每一个trait都有一个默认的编辑器，将管理trait在屏幕上显示的方式 (即Range trait显示为一个滑块等)。

与Traits声明方式来创建类的相同渠道，TraitsUI提供了声明的界面来构建用户界面代码:

In [ ]:

  1. from traits.api import HasTraits, Str, Float, Range
  2. from traitsui.api import View
  4. class Reservoir(HasTraits):
  5.     name = Str
  6.     max_storage = Float(1e6, desc='Maximal storage [hm3]')
  7.     max_release = Float(10, desc='Maximal release [m3/s]')
  8.     head = Float(10, desc='Hydraulic head [m]')
  9.     efficiency = Range(0, 1.)
  11.     traits_view = View(
  12.         'name', 'max_storage', 'max_release', 'head', 'efficiency',
  13.         title = 'Reservoir',
  14.         resizable = True,
  15.     )
  17.     def energy_production(self, release):
  18.         ''' Returns the energy production [Wh] for the given release [m3/s]
  19.  '''
  20.         power = 1000 * 9.81 * self.head * release * self.efficiency 
  21.         return power * 3600
  23. if __name__ == '__main__':
  24.     reservoir = Reservoir(
  25.                         name = 'Project A',
  26.                         max_storage = 30,
  27.                         max_release = 100.0,
  28.                         head = 60,
  29.                         efficiency = 0.8
  30.                     )
  32.     reservoir.configure_traits()

3.4.3 Traits是什么 - 图2

3.4.3.5 推迟

可以将trait定义和它的值推送给另一个对象是Traits的有用的功能。

In [ ]:

  1. from traits.api import HasTraits, Instance, DelegatesTo, Float, Range
  3. from reservoir import Reservoir
  5. class ReservoirState(HasTraits):
  6.     """Keeps track of the reservoir state given the initial storage.
  7.  """
  8.     reservoir = Instance(Reservoir, ())
  9.     min_storage = Float
  10.     max_storage = DelegatesTo('reservoir')
  11.     min_release = Float
  12.     max_release = DelegatesTo('reservoir')
  14.     # state attributes
  15.     storage = Range(low='min_storage', high='max_storage')
  17.     # control attributes
  18.     inflows =  Float(desc='Inflows [hm3]')
  19.     release = Range(low='min_release', high='max_release')
  20.     spillage = Float(desc='Spillage [hm3]')
  22.     def print_state(self):
  23.         print 'Storage\tRelease\tInflows\tSpillage'
  24.         str_format = '\t'.join(['{:7.2f}'for i in range(4)])
  25.         print str_format.format(self.storage, self.release, self.inflows,
  26.                 self.spillage)
  27.         print '-' * 79
  29. if __name__ == '__main__':
  30.     projectA = Reservoir(
  31.             name = 'Project A',
  32.             max_storage = 30,
  33.             max_release = 100.0,
  34.             hydraulic_head = 60,
  35.             efficiency = 0.8
  36.         )
  38.     state = ReservoirState(reservoir=projectA, storage=10)
  39.     state.release = 90
  40.     state.inflows = 0
  41.     state.print_state()
  43.     print 'How do we update the current storage ?'

特殊的trait允许用魔法_xxxx_fired方法管理事件和触发器函数:

In [ ]:

  1. from traits.api import HasTraits, Instance, DelegatesTo, Float, Range, Event
  3. from reservoir import Reservoir
  5. class ReservoirState(HasTraits):
  6.     """Keeps track of the reservoir state given the initial storage.
  8.  For the simplicity of the example, the release is considered in
  9.  hm3/timestep and not in m3/s.
  10.  """
  11.     reservoir = Instance(Reservoir, ())
  12.     min_storage = Float
  13.     max_storage = DelegatesTo('reservoir')
  14.     min_release = Float
  15.     max_release = DelegatesTo('reservoir')
  17.     # state attributes
  18.     storage = Range(low='min_storage', high='max_storage')
  20.     # control attributes
  21.     inflows =  Float(desc='Inflows [hm3]')
  22.     release = Range(low='min_release', high='max_release')
  23.     spillage = Float(desc='Spillage [hm3]')
  25.     update_storage = Event(desc='Updates the storage to the next time step')
  27.     def _update_storage_fired(self):
  28.         # update storage state
  29.         new_storage = self.storage - self.release  + self.inflows
  30.         self.storage = min(new_storage, self.max_storage)
  31.         overflow = new_storage - self.max_storage
  32.         self.spillage = max(overflow, 0)
  34.     def print_state(self):
  35.         print 'Storage\tRelease\tInflows\tSpillage'
  36.         str_format = '\t'.join(['{:7.2f}'for i in range(4)])
  37.         print str_format.format(self.storage, self.release, self.inflows,
  38.                 self.spillage)
  39.         print '-' * 79
  41. if __name__ == '__main__':
  42.     projectA = Reservoir(
  43.         name = 'Project A',
  44.         max_storage = 30,
  45.         max_release = 5.0,
  46.         hydraulic_head = 60,
  47.         efficiency = 0.8
  48.     )
  50.     state = ReservoirState(reservoir=projectA, storage=15)
  51.     state.release = 5
  52.     state.inflows = 0
  54.     # release the maximum amount of water during 3 time steps
  55.     state.update_storage = True
  56.     state.print_state()
  57.     state.update_storage = True
  58.     state.print_state()
  59.     state.update_storage = True
  60.     state.print_state()

对象间的依赖可以自动使用traitProperty完成。depends_on属性表示property其他traits的依赖性。当其他traits改变了,property是无效的。此外，Traits为属性使用魔法函数的名字:

  • _get_XXX 来获得XXX属性的trait
  • _set_XXX 来设置XXX属性的trait

In [ ]:

  1. from traits.api import HasTraits, Instance, DelegatesTo, Float, Range
  2. from traits.api import Property
  4. from reservoir import Reservoir
  6. class ReservoirState(HasTraits):
  7.     """Keeps track of the reservoir state given the initial storage.
  9.  For the simplicity of the example, the release is considered in
  10.  hm3/timestep and not in m3/s.
  11.  """
  12.     reservoir = Instance(Reservoir, ())
  13.     max_storage = DelegatesTo('reservoir')
  14.     min_release = Float
  15.     max_release = DelegatesTo('reservoir')
  17.     # state attributes
  18.     storage = Property(depends_on='inflows, release')
  20.     # control attributes
  21.     inflows =  Float(desc='Inflows [hm3]')
  22.     release = Range(low='min_release', high='max_release')
  23.     spillage = Property(
  24.             desc='Spillage [hm3]', depends_on=['storage', 'inflows', 'release']
  25.         )
  27.     ### Private traits.
  28.     _storage = Float
  30.     ### Traits property implementation.
  31.     def _get_storage(self):
  32.         new_storage = self._storage - self.release + self.inflows
  33.         return min(new_storage, self.max_storage)
  35.     def _set_storage(self, storage_value):
  36.         self._storage = storage_value
  38.     def _get_spillage(self):
  39.         new_storage = self._storage - self.release  + self.inflows
  40.         overflow = new_storage - self.max_storage
  41.         return max(overflow, 0)
  43.     def print_state(self):
  44.         print 'Storage\tRelease\tInflows\tSpillage'
  45.         str_format = '\t'.join(['{:7.2f}'for i in range(4)])
  46.         print str_format.format(self.storage, self.release, self.inflows,
  47.                 self.spillage)
  48.         print '-' * 79
  50. if __name__ == '__main__':
  51.     projectA = Reservoir(
  52.                     name = 'Project A',
  53.                     max_storage = 30,
  54.                     max_release = 5,
  55.                     hydraulic_head = 60,
  56.                     efficiency = 0.8
  57.                 )
  59.     state = ReservoirState(reservoir=projectA, storage=25)
  60.     state.release = 4
  61.     state.inflows = 0
  63.     state.print_state()

注意 缓存属性 当访问一个输入没有改变的属性时，[email protected]_property修饰器可以用来缓存这个值，并且只有在失效时才会重新计算一次他们。

让我们用ReservoirState的例子来扩展TraitsUI介绍:

In [ ]:

  1. from traits.api import HasTraits, Instance, DelegatesTo, Float, Range, Property
  2. from traitsui.api import View, Item, Group, VGroup
  4. from reservoir import Reservoir
  6. class ReservoirState(HasTraits):
  7.     """Keeps track of the reservoir state given the initial storage.
  9.  For the simplicity of the example, the release is considered in
  10.  hm3/timestep and not in m3/s.
  11.  """
  12.     reservoir = Instance(Reservoir, ())
  13.     name = DelegatesTo('reservoir')
  14.     max_storage = DelegatesTo('reservoir')
  15.     max_release = DelegatesTo('reservoir')
  16.     min_release = Float
  18.     # state attributes
  19.     storage = Property(depends_on='inflows, release')
  21.     # control attributes
  22.     inflows =  Float(desc='Inflows [hm3]')
  23.     release = Range(low='min_release', high='max_release')
  24.     spillage = Property(
  25.             desc='Spillage [hm3]', depends_on=['storage', 'inflows', 'release']
  26.         )
  28.     ### Traits view
  29.     traits_view = View(
  30.         Group(
  31.             VGroup(Item('name'), Item('storage'), Item('spillage'),
  32.                 label = 'State', style = 'readonly'
  33.             ),
  34.             VGroup(Item('inflows'), Item('release'), label='Control'),
  35.         )
  36.     )
  38.     ### Private traits.
  39.     _storage = Float
  41.     ### Traits property implementation.
  42.     def _get_storage(self):
  43.         new_storage = self._storage - self.release + self.inflows
  44.         return min(new_storage, self.max_storage)
  46.     def _set_storage(self, storage_value):
  47.         self._storage = storage_value
  49.     def _get_spillage(self):
  50.         new_storage = self._storage - self.release  + self.inflows
  51.         overflow = new_storage - self.max_storage
  52.         return max(overflow, 0)
  54.     def print_state(self):
  55.         print 'Storage\tRelease\tInflows\tSpillage'
  56.         str_format = '\t'.join(['{:7.2f}'for i in range(4)])
  57.         print str_format.format(self.storage, self.release, self.inflows,
  58.                 self.spillage)
  59.         print '-' * 79
  61. if __name__ == '__main__':
  62.     projectA = Reservoir(
  63.         name = 'Project A',
  64.         max_storage = 30,
  65.         max_release = 5,
  66.         hydraulic_head = 60,
  67.         efficiency = 0.8
  68.     )
  70.     state = ReservoirState(reservoir=projectA, storage=25)
  71.     state.release = 4
  72.     state.inflows = 0
  74.     state.print_state()
  75.     state.configure_traits()

3.4.3 Traits是什么 - 图3

Some use cases need the delegation mechanism to be broken by the user when setting the value of the trait. The PrototypeFrom trait implements this behaviour.

In [ ]:

  1. TraitsUI simplifies the way user interfaces are created. Every trait on a HasTraits class has a default editor that will manage the way the trait is rendered to the screen (e.g. the Range trait is displayed as a slider, etc.).
  2. In the very same vein as the Traits declarative way of creating classes, TraitsUI provides a declarative interface to build user interfaces code: