def compute(self):
"""
构建时间序列变量，执行指标的向量算法。
"""
if not hasattr(self, '_args'):
raise Exception("每个指标都必须有_args属性，代表指标计算的参数！")
self.data = self._args[0]
# 数据转化成ta-lib能处理的格式
#self._args[0] = ndarray(self._args[0])
apply(self._vector_algo, tuple(self._args))
if not hasattr(self, 'values'):
raise Exception("每个指标都必须有value属性，代表指标计算结果！")
if isinstance(self.values, dict):
self.series = OrderedDict()
for key, value in self.values.iteritems():
self.series[key] = series.NumberSeries(
value, self.name, self, float('nan'))
for key, value in self.series.iteritems():
setattr(self, key, value)
self.is_multiple = True
else:
self.series = [series.NumberSeries(
self.values, self.name, self, float('nan'))]
self.is_multiple = False
self._init_bound()
def compute(self):
"""
构建时间序列变量，执行指标的向量算法。
"""
if not hasattr(self, '_args'):
raise Exception("每个指标都必须有_args属性，代表指标计算的参数！")
self.data = self._args[0]
if not series.g_rolling:
# 数据转化成ta-lib能处理的格式
self._args[0] = transform2ndarray(self._args[0])
apply(self._vector_algo, tuple(self._args))
if not hasattr(self, 'values'):
raise Exception("每个指标都必须有value属性，代表指标计算结果！")
if isinstance(self.values, dict):
self.series = OrderedDict()
for key, value in self.values.iteritems():
self.series[key] = series.NumberSeries(value, len(value),
self.name, self, float('nan'))
for key, value in self.series.iteritems():
setattr(self, key, value)
self.multi_value = True
else:
self.series = [series.NumberSeries(self.values, len(self.values),
self.name, self, float('nan'))]
self.multi_value = False
if series.g_rolling:
if self.multi_value:
for key, value in self.series.iteritems():
value.reset_data([], 1+series.g_window)
else:
for s in self.series:
s.reset_data([], 1+series.g_window)
self._init_bound()
def wrapper(self, *args, **kwargs):
magic = inspect.getargspec(method)
arg_names = magic.args[1:]
# 默认参数
default = dict(
(x, y)
for x, y in zip(magic.args[-len(magic.defaults):], magic.defaults))
# 调用参数
method_args = {}
for i, arg in enumerate(args):
method_args[arg_names[i]] = arg
method_args.update(kwargs)
#
default.update(method_args)
# 属性创建
for key, value in default.iteritems():
setattr(self, key, value)
# 构造函数
rst = method(self, *args, **kwargs)
if not hasattr(self, 'value'):
raise Exception("每个指标都必须有value属性，代表指标值！")
else:
# 序列变量
if self.tracker:
if isinstance(self.value, tuple):
self._series = [
series.NumberSeries(self.tracker, value)
for value in self.value
]
else:
self._series = series.NumberSeries(self.tracker,
self.value)
# 绘图中的y轴范围未被设置，使用默认值。
if not self.upper:
upper = lower = []
if isinstance(self.value, tuple):
# 多值指标
upper = [
max([value[i] for value in self.value])
for i in xrange(0, len(self.value[0]))
]
lower = [
min([value[i] for value in self.value])
for i in xrange(0, len(self.value[0]))
]
else:
upper = self.value
lower = self.value
self.set_yrange(lower, upper)
return rst
def _init_by_subclass(self, data):
# 数据转化成ta-lib能处理的格式
# self.value为任何支持index的数据结构。
# 在策略中，price变量可能为NumberSeries，需要用NUMBER_SERIES_SUPPORT处理，
# 转化为numpy.ndarray等能被指标函数处理的参数。
if not series.g_rolling:
# 向量化运行的均值函数
data = transform2ndarray(data)
args = (data, ) + self._args
apply(self._vector_algo, args)
#if not hasattr(self, 'value'):
#raise Exception("每个指标都必须有value属性，代表指标值！")
#else:
if isinstance(self.value, dict):
self.series = OrderedDict()
for key, value in self.value.iteritems():
self.series[key] = series.NumberSeries(value, len(value),
self.name, self, float('nan'))
for key, value in self.series.iteritems():
setattr(self, key, value)
self.multi_value = True
else:
self.series = [series.NumberSeries(self.value, len(self.value),
self.name, self, float('nan'))]
self.multi_value = False
if series.g_rolling:
if self.multi_value:
for key, value in self.series.iteritems():
value.reset_data([], 1+series.g_window)
else:
for s in self.series:
s.reset_data([], 1+series.g_window)
# 绘图中的y轴范围未被设置，使用默认值。
if not self._upper:
upper = lower = []
if isinstance(self.value, tuple):
# 多值指标
upper = [ max([value[i] for value in self.value ])
for i in xrange(0, len(self.value[0]))]
lower = [ min([value[i] for value in self.value ])
for i in xrange(0, len(self.value[0]))]
else:
upper = self.value
lower = self.value
self.set_yrange(lower, upper)
