[python] A note that started to understand the behavior of matplotlib.pyplot

Pyplot, a module for drawing graphs, is sometimes taken care of, but I was not accustomed to using it. I've finally been able to eliminate that haze recently, so I'll leave it as a note of my awareness.

• The content is not "why it is designed like that" but "what kind of design it is".

Where I didn't understand personally

First is the basic graph output.

import matplotlib.pyplot as plt
import numpy as np

x = np.linspace(-np.pi, np.pi)
y = np.sin(x)

fig = plt.figure()
plt.plot(x, y)
plt.title("y = sin(x)")
plt.xlabel("x")
plt.ylabel("y")
fig.savefig("output.png ")

With this, the image is output as ʻoutput.png`, but personally, the behavior at this time did not fall into my mind. For example

--There is no evidence of creating a plt instance --I used the plt method to change the title of the graph, but I called the fig method to save it as an image. --The target when accessing with plt just by generating fig1, fig2, ... like fig changes in order as fig1, fig2, ...

(By the way, I didn't think that title () and xlabel () are class methods). Personally, I'm used to the flow of instance creation → method call, so I think I understood without hesitation if the method was as below.

#This does not work properly.
import matplotlib
import numpy as np

x = np.linspace(-np.pi, np.pi)
y = np.sin(x)

plt = matplotlib.pyplot() #I want you to create an instance here
plt.plot(x, y)
plt.title("y = sin(x)")
plt.xlabel("x")
plt.ylabel("y")
plt.figure.savefig("output.png ") #I want this to be a method of the plt instance

Why is it successful in the first way? And why doesn't it work the way I envisioned?

What's really going on

Find out who each of the elements in the first working script is (omip the numpy part).

• matplotlib --Library for drawing graphs
• pyplot --Modules in the matplotlib library --ʻImport Import and use with matplotlib.pyplot` etc.
• plt --Alias for matplotlib.pyplot (hence type is module) --Since it is an alias, you can change it to any name you like.
• fig --An instance of the matplotlib.figure.Figure class --Generated by fig = pyplot.figure ()

The important thing is that ** plt is a module **, not a class. If plt was a class, I would like it to be instantiated somewhere, but because it was a module, it was already created when I imported it on the second line, so suddenlyfig = plt.figure ()andIt worked fine with plt.plot (x, y). Also, this is speculation, but if you think that the access destination is updated in plt when the fig instance is created, other questions may be convincing.

bonus

Now, in python, modules are created as singletons, so they refer to objects with the same id no matter where you call them. Therefore, if you call pyplot elsewhere, you will still be able to access the graph you have already created.

For example, suppose the following two files are in the same directory:

graph.py

import matplotlib.pyplot as plt
import numpy as np

def edit_graph():
    x = np.linspace(-np.pi, np.pi)
    y = np.sin(x)

    plt.plot(x, y)
    plt.title("y = sin(x)")
    plt.xlabel("x")
    plt.ylabel("y")

main.py

import matplotlib.pyplot as plt_main
import graph

graph.edit_graph()
plt_main.show()

If you execute python main.py in this state, the same sine wave graph as before will be output. In other words, it is possible to output a graph generated / edited in one script file in another script file without creating / passing an instance (although it is not recommended by design). You can also call the method savefig () in the matplotlib.figure.Figure class after executingfig = pyplot.figure ().

Summary

The main reason for this question was that I misunderstood plt as some kind of class or instance. I don't know why I thought so now ... It's not a big deal if you realize that plt is a module, but I think it will be better understood if you think about it again. ..

Finally, a summary of what I noticed this time.

--pyplot is a module and is usually used with a name like plt. --The pyplot module itself behaves like a class or its instance --Every time you execute fig = pyplot.figure (), a fig instance of the matplotlib.figure.Figure class is created, and each time you execute it, the access target by the pyplot module changes.