本文整理汇总了Python中geographiclib.geomath.Math.sum方法的典型用法代码示例。如果您正苦于以下问题：Python Math.sum方法的具体用法？Python Math.sum怎么用？Python Math.sum使用的例子？那么恭喜您, 这里精选的方法代码示例或许可以为您提供帮助。您也可以进一步了解该方法所在类geographiclib.geomath.Math的用法示例。
在下文中一共展示了Math.sum方法的1个代码示例，这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞，您的评价将有助于系统推荐出更棒的Python代码示例。
示例1: Add
# 需要导入模块: from geographiclib.geomath import Math [as 别名]
# 或者: from geographiclib.geomath.Math import sum [as 别名]
def Add(self, y):
"""Add a value"""
# Here's Shewchuk's solution...
# hold exact sum as [s, t, u]
y, u = Math.sum(y, self._t) # Accumulate starting at
self._s, self._t = Math.sum(y, self._s) # least significant end
# Start is _s, _t decreasing and non-adjacent. Sum is now (s + t + u)
# exactly with s, t, u non-adjacent and in decreasing order (except
# for possible zeros). The following code tries to normalize the
# result. Ideally, we want _s = round(s+t+u) and _u = round(s+t+u -
# _s). The follow does an approximate job (and maintains the
# decreasing non-adjacent property). Here are two "failures" using
# 3-bit floats:
#
# Case 1: _s is not equal to round(s+t+u) -- off by 1 ulp
# [12, -1] - 8 -> [4, 0, -1] -> [4, -1] = 3 should be [3, 0] = 3
#
# Case 2: _s+_t is not as close to s+t+u as it shold be
# [64, 5] + 4 -> [64, 8, 1] -> [64, 8] = 72 (off by 1)
# should be [80, -7] = 73 (exact)
#
# "Fixing" these problems is probably not worth the expense. The
# representation inevitably leads to small errors in the accumulated
# values. The additional errors illustrated here amount to 1 ulp of
# the less significant word during each addition to the Accumulator
# and an additional possible error of 1 ulp in the reported sum.
#
# Incidentally, the "ideal" representation described above is not
# canonical, because _s = round(_s + _t) may not be true. For
# example, with 3-bit floats:
#
# [128, 16] + 1 -> [160, -16] -- 160 = round(145).
# But [160, 0] - 16 -> [128, 16] -- 128 = round(144).
#
if self._s == 0: # This implies t == 0,
self._s = u # so result is u
else:
self._t += u # otherwise just accumulate u to t.