使用matplotlib / pyplot绘制图形并仅保存图形（居中，长宽比相同且没有轴或边框）

Question

I want to save just the resulting plot image to a file and wrote the following MCVE:

import matplotlib.pyplot as plt

plt.rcParams["axes.facecolor"] = "red"
plt.rcParams["savefig.facecolor"] = "red"
plt.rcParams["savefig.bbox"] = "tight"
plt.rcParams["savefig.pad_inches"] = 0

plt.plot([0, 1], [0, 0], "b-", linewidth="10")
plt.plot([1, 1], [0, 1], "b-", linewidth="10")
plt.plot([1, 0], [1, 1], "b-", linewidth="10")
plt.plot([0, 0], [1, 0], "b-", linewidth="10")

#plt.axis("equal")
#plt.axis("off")
plt.savefig("test.png")

---

plt.axis("equal")
#plt.axis("off")

---

#plt.axis("equal")
plt.axis("off")

---

plt.axis("equal")
plt.axis("off")

---

I found out about bbox and pad_inches . However, it's still not perfect. The axes seem to be hidden, not off completely - so the plot won't be centered. Additionally, I think I even need the axes to set a equal ratio or the rectangle won't be a square?

What I want is

or

Answer 1

You can define a figure in a square shape and use tight_layout() to make the padding compact automatically.

import matplotlib.pyplot as plt

# Creates a figure object of size 6x6 inches
fig = plt.figure(figsize=(6,6))

plt.rcParams["axes.facecolor"] = "red"
plt.rcParams["savefig.facecolor"] = "red"
plt.rcParams["savefig.bbox"] = "tight"
plt.rcParams["savefig.pad_inches"] = 0

plt.plot([0, 1], [0, 0], "b-", linewidth="10")
plt.plot([1, 1], [0, 1], "b-", linewidth="10")
plt.plot([1, 0], [1, 1], "b-", linewidth="10")
plt.plot([0, 0], [1, 0], "b-", linewidth="10")

plt.axis("equal")
plt.axis("off")
fig.tight_layout()
fig.savefig("test.png")

Answer 2

plt.rcParams["savefig.bbox"] = "tight" adjusts the padding around the figure automatically. This seems to be undesired here. Leaving that out and creating a square figure with no margins ( plt.subplots_adjust(0,0,1,1) ) will allow to have the axes take all the space in the figure. There is then no need to set the aspect specifically any more, but of course for different cases that may still make sense.

import matplotlib.pyplot as plt

plt.rcParams["axes.facecolor"] = "red"
plt.rcParams["savefig.facecolor"] = "red"

plt.figure(figsize=(4,4))
plt.subplots_adjust(0,0,1,1)
plt.plot([0, 1, 1, 0,0], [0, 0,1,1,0], "b-", linewidth="10", solid_joinstyle="miter")

#plt.gca().set_aspect("equal",adjustable="box" ) # <- not needed
plt.axis("off")
plt.savefig("test.png")
plt.show()

To obtain the figure without red border, you could set the data margins to 0% ( plt.margins(0) ); however, because the line is half cut, it makes sense to double its linewidth to obtain the same blue border.

import matplotlib.pyplot as plt

plt.rcParams["axes.facecolor"] = "red"
plt.rcParams["savefig.facecolor"] = "red"

plt.figure(figsize=(4,4))
plt.subplots_adjust(0,0,1,1)
plt.plot([0, 1, 1, 0,0], [0, 0,1,1,0], "b-", linewidth=20, solid_joinstyle="miter")

#plt.gca().set_aspect("equal",adjustable="box" ) # <- not needed
plt.axis("off")
plt.margins(0.0)
plt.savefig("test.png")
plt.show()