在matplotlib图上嵌入小地图（cartopy）

Question

# imports
from collections import namedtuple
import numpy as np
import xarray as xr
import shapely
import cartopy

 The data that I have looks like this.

我有一个感兴趣的区域（在这里定义为all_region）。我有一个xr.DataArray，其中包含我的变量。我想做的是选择一个PIXEL（lat，lon对）并在线图的角落绘制一个小地图，显示像素所在的位置。

Region = namedtuple('Region',field_names=['region_name','lonmin','lonmax','latmin','latmax'])
all_region = Region(
    region_name="all_region",
    lonmin = 32.6,
    lonmax = 51.8,
    latmin = -5.0,
    latmax = 15.2,
)

data = np.random.normal(0,1,(12, 414, 395))
lats = np.linspace(-4.909738, 15.155708, 414)
lons = np.linspace(32.605801, 51.794488, 395)
months = np.arange(1,13)
da = xr.DataArray(data, coords=[months, lats, lons], dims=['month','lat','lon'])

These are the functions that I need to fix to work with inset axes.

我有这些函数从xarray对象绘制我的时间序列，以及lat，lon点的位置。

def plot_location(region):
    """ use cartopy to plot the region (defined as a namedtuple object)
    """
    lonmin,lonmax,latmin,latmax = region.lonmin,region.lonmax,region.latmin,region.latmax
    fig = plt.figure()
    ax = fig.gca(projection=cartopy.crs.PlateCarree())
    ax.add_feature(cartopy.feature.COASTLINE)
    ax.add_feature(cartopy.feature.BORDERS, linestyle=':')
    ax.set_extent([lonmin, lonmax, latmin, latmax])

    return fig, ax


def select_pixel(ds, loc):
    """ (lat,lon) """
    return ds.sel(lat=loc[1],lon=loc[0],method='nearest')


def turn_tuple_to_point(loc):
    """ (lat,lon) """
    from shapely.geometry.point import Point
    point = Point(loc[1], loc[0])
    return point


def add_point_location_to_map(point, ax, color=(0,0,0,1), **kwargs):
    """ """
    ax.scatter(point.x,
           point.y,
           transform=cartopy.crs.PlateCarree(),
           c=[color],
           **kwargs)
    return

Here I do the plotting

# choose a lat lon location that want to plot
loc = (2.407,38.1)

# 1. plot the TIME SERIES FOR THE POINT
fig,ax = plt.subplots()
pixel_da = select_pixel(da, loc)
pixel_da.plot.line(ax=ax, marker='o')

# 2. plot the LOCATION for the point
fig,ax = plot_location(all_region)
point = turn_tuple_to_point(loc)
add_point_location_to_map(point, ax)

 I have my function for plotting a region, but I want to put this on an axis in the corner of my figure! Like this:

我该怎么做呢？我看过inset_locator method，但据我所知，mpl_toolkits.axes_grid1.parasite_axes.AxesHostAxes无法分配投影，这是投票所必需的。

from mpl_toolkits.axes_grid1.inset_locator import inset_axes

proj=cartopy.crs.PlateCarree
axins = inset_axes(ax, width="20%", height="20%", loc=2, projection=proj)

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-162-9b5fd4f34c3e> in <module>
----> 1 axins = inset_axes(ax, width="20%", height="20%", loc=2, projection=proj)

TypeError: inset_axes() got an unexpected keyword argument 'projection'

Answer 1

mpl_toolkits.axes_grid1.inset_locator.inset_axes没有projection关键字。它只提供了axes_class论证。现在有人可能想要将cartopy.mpl.geoaxes.GeoAxes直接提供给该参数，但这将缺少使用中的实际投影。所以另外需要通过axes_kwargs参数设置投影。

inset_axes(...,  axes_class=cartopy.mpl.geoaxes.GeoAxes, 
                 axes_kwargs=dict(map_projection=cartopy.crs.PlateCarree()))

完整的例子：

import cartopy
import cartopy.mpl.geoaxes
import matplotlib.pyplot as plt
from mpl_toolkits.axes_grid1.inset_locator import inset_axes


fig, ax = plt.subplots()
ax.plot([4,5,3,1,2])


axins = inset_axes(ax, width="40%", height="40%", loc="upper right", 
                   axes_class=cartopy.mpl.geoaxes.GeoAxes, 
                   axes_kwargs=dict(map_projection=cartopy.crs.PlateCarree()))
axins.add_feature(cartopy.feature.COASTLINE)


plt.show()

Answer 2

我没有安装cartopy直接测试它，但我相信你可以通过直接手工创建你的插入轴来解决问题，using fig.add_axes()。如果要指定其相对于主轴的位置，可以使用主轴rect返回的信息轻松计算get_position()参数。

例如：

pad = 0.05
w = 0.4
h = 0.25

fig, ax = plt.subplots()
a = ax.get_position()
ax2 = fig.add_axes([a.x1-(w+pad)*a.width,a.y1-(h+pad)*a.height,w*a.width,h*a.height], projection="hammer")