旋转FancyArrowPatch

Question

我正在尝试将FancyArrowPatch添加到子图，并根据度数顺序旋转图块。我尝试了变换的几种变体，但是没有一个可行的解决方案。 This solution令我赞叹，好像它应该可以工作，但是当我添加转换时，补丁消失了（转换运行没有错误）。我也尝试过mpl.patches.arrow()。

我不愿意使用FancyArrowPatch-任何形式的方向标记都足够-只要它们可以独立旋转以反映风向即可。我仅限于matplotlib 1.3.1；在这种情况下，不能选择迁移到新版本。我强烈怀疑这与以下事实有关：x轴是基于时间的，并且/或者由于收集而涉及某种方式的混乱，但这只是一个猜测。

# -*- coding: utf-8 -*-

import datetime as dt
import matplotlib.pyplot as plt  # vers 1.3.1
import matplotlib.dates as md
import matplotlib.patches as mpatches
import matplotlib.transforms as transform

x_offset = dt.timedelta(hours=4)

fig, (ax0, ax1, ax2) = plt.subplots(nrows=3, sharex=True)

x0 = [1558065600, 1558152000, 1558238400, 1558324800, 1558411200, 1558497600, 1558584000, 1558670400]  # <-- Date
x1 = [dt.datetime.fromtimestamp(_) for _ in x0]

# ================================= Subplot 1 =================================
y  = [83.12, 82.45, 87.51, 85.27, 75.85, 81.72, 91.16, 86.74]  # <-- Temperature

ax0.set_title('high temperature')
ax0.plot(x1, y)
ax0.set_xlim(min(x1) - x_offset, max(x1) + x_offset)
myFmt = md.DateFormatter('%m-%d')
ax1.xaxis.set_major_formatter(myFmt)

# ================================= Subplot 2 =================================
y    = [4.25, 2.03, 7.77, 9.85, 6.73, 4.47, 5.97, 7.24]  # <-- Wind Speed
z    = [251, 86, 171, 224, 331, 147, 265, 319]  # <-- Wind Bearing
data = zip(x1, y, z)

ax1.set_title('wind')
ax1.plot(x1, y)
ax1.set_ylim(0, max(y) + 1)

for _ in data:
    day      = md.date2num(_[0])
    location = _[1]
    bearing  = _[2]  #  <-- Rotate patch based on this

    arr = mpatches.FancyArrowPatch(posA=(day, location), posB=(day, location + 0.1), arrowstyle='Simple', color='red', lw=0)
    arr.set_arrowstyle("fancy", head_length=10, head_width=10)

    # <-- Perform elusive transformation -->

    ax1.add_patch(arr)

# ================================= Subplot 3 =================================
# Precip intensity is in inches of liquid rain per hour.
y  = [_ * 24 for _ in [0.0021, 0.0035, 0.0018, 0.0017, 0.0001, 0.0000, 0.0001, 0.0003]]  # <-- Precip

ax2.set_title('precip intensity')
ax2.plot(x1, y)

plt.show()

更新

基于Sheldore的建议（示例1），我能够找到解决方案。新代码的关键部分取代了对FancyArrowPatch的调用。

# -*- coding: utf-8 -*-

import datetime as dt
import matplotlib.pyplot as plt  # vers 1.3.1
import matplotlib.dates as md
import matplotlib.patches as mpatches
import matplotlib.transforms as transform

x_offset = dt.timedelta(hours=4)

fig, (ax0, ax1, ax2) = plt.subplots(nrows=3, sharex=True)

x0 = [1558065600, 1558152000, 1558238400, 1558324800, 1558411200, 1558497600, 1558584000, 1558670400]  # <-- Date
x1 = [dt.datetime.fromtimestamp(_) for _ in x0]

# ================================= Subplot 1 =================================
y  = [83.12, 82.45, 87.51, 85.27, 75.85, 81.72, 91.16, 86.74]  # <-- Temperature

ax0.set_title('high temperature')
ax0.plot(x1, y)
ax0.set_xlim(min(x1) - x_offset, max(x1) + x_offset)
myFmt = md.DateFormatter('%m-%d')
ax1.xaxis.set_major_formatter(myFmt)

# ================================= Subplot 2 =================================
y    = [4.25, 2.03, 7.77, 9.85, 6.73, 4.47, 5.97, 7.24]  # <-- Wind Speed
z    = [251, 86, 171, 224, 331, 147, 265, 319]  # <-- Wind Bearing
data = zip(x1, y, z)

ax1.set_title('wind')
ax1.plot(x1, y)
ax1.set_ylim(0, max(y) + 1)

for _ in data:
    day      = md.date2num(_[0])
    location = _[1]

    # New code to replace FancyArrowPatch
    ax1.annotate('', xy=(day, location), xycoords='data', xytext=(0, 1), textcoords='offset points', rotation=0, size=20,
                 arrowprops=dict(arrowstyle="simple", fc="red", ec="none", connectionstyle="angle3,angleA=0,angleB={0}".format(_[2])))

# ================================= Subplot 3 =================================
# Precip intensity is in inches of liquid rain per hour.
y  = [_ * 24 for _ in [0.0021, 0.0035, 0.0018, 0.0017, 0.0001, 0.0000, 0.0001, 0.0003]]  # <-- Precip

ax2.set_title('precip intensity')
ax2.plot(x1, y)

plt.show()

我需要稍微清理一下，但是解决方案会产生预期的结果。