Question

如何使用Matplotlib的Gridspec来拟合/调整大小/屏蔽图像以适合（精确）在列单元格内？

下面的代码需要一个软糖因子53来调整图像大小以适合（确切地）在单元格内。

import matplotlib.gridspec as gridspec
from matplotlib.offsetbox import OffsetImage, AnnotationBbox
%config InlineBackend.figure_formats = {'png', 'retina'}
%matplotlib inline

side = 10 
nrows=5 
ncols =5
fudge = 53

fig = plt.figure(figsize=(15, 15))
gs = gridspec.GridSpec(nrows=5, ncols=5, wspace=0, hspace=0, 
                       width_ratios=[1,1,1,1,1], 
                       height_ratios=[1,1,1,1,1])

axes = {}

######################
# Thumbnail Image Column

for i in range(0,nrows):
    axes[i] = plt.subplot(gs[i,0])
    axes[i].set_xlim(0,10)
    axes[i].set_ylim(0,10)
    axes[i].axis("on")

    # Add Grey Square
    square = plt.Rectangle((0, 0), side, side, fc='b', hatch='///', linewidth=1, color='grey' )
    axes[i].add_patch(square)

    bbox = axes[i].get_window_extent().transformed(fig.dpi_scale_trans.inverted())
    width, height = bbox.width, bbox.height
    width *= fig.dpi *2
    height *= fig.dpi*2

    print(width,height, fig.dpi)
    #print(fig.dpi, bbox.width, width, bbox.height, height)

    # Trying to Mask/Resize the Image so it fits within the Column Cell
    img = Image.open("./images/test_500x500_pixel.png")
    #img_crop = img.crop((0, 0 ,width, height)) # Not using crop

    img.thumbnail((width+fudge, height+fudge), Image.ANTIALIAS)
    imagebox = OffsetImage(img, zoom=1, alpha=1)
    ab = AnnotationBbox(imagebox, xy=(0,0),
                        xycoords='data',
                        frameon=False,
                        box_alignment=(0,0),
                        boxcoords = None,
                        pad=0.0  
                       )
    axes[i].add_artist(ab)  
plt.tight_layout(pad=0)

我试图让图像适合列单元格 - 无论图形大小如何。此列包含缩略图图像。其他列是从python pandas数据帧中提取的数据。

非常感谢提前。

Answer 1

以下非常简单的代码为image列提供了相同的宽高比。

import numpy as np
import matplotlib.pyplot as plt

image = plt.imread("https://i.stack.imgur.com/ltBO1.png")

N = 5
fig, ax = plt.subplots(N,2, figsize=(5,8))

for i in range(N):
    thisimage = np.copy(image)
    thisimage[:,:,0] = thisimage[:,:,0]*(i+1)/float(N)
    ax[i,0].imshow(thisimage)
    ax[i,0].set_xticks([])
    ax[i,0].set_yticks([])
    ax[i,1].plot(np.random.rand(9))

plt.show()

它产生以下图表。

我不清楚这有多远不符合要求？

调整图像大小以完全适合（Matplotlib）GridSpec轴

1 个答案: