算法的动画可视化

Question

我想知道是否有一种方法可以创建漂亮的可视化（在 Python 中），比如涉及图形的算法。

如果有一种方法可以在 Python 中做到这一点，这将有助于将算法代码的每个执行逻辑步骤转换为简洁的实时插图。

在 Wikipedia 上阅读 TSP 时，我发现了这一点：

Answer 1

我一直使用从 matplotlib 创建的单个图来做这件事。

一个示例程序是：

1. 创建多个绘图并将它们保存为图像文件
2. 遍历每个保存的图像文件并使用opencv 读取它们
3. 使用opencv 将所有图像文件编译成一个视频文件。

这是一些简化的示例代码

import cv2
import os
import matplotlib.pyplot as plt

# create a single plot
plt.plot([1,2,3], [3, 7, 11])
# save plot as an image
plt.savefig(plot_directory\plot_name.jpg, format='jpg', dpi=250)
plt.show()


def create_video(image_folder, video_name, fps=8, reverse=False):
    """Create video out of images saved in a folder."""
    images = [img for img in os.listdir(image_folder) if img.endswith('.jpg')]
    if reverse: images = images[::-1]
    frame = cv2.imread(os.path.join(image_folder, images[0]))
    height, width, layers = frame.shape
    video = cv2.VideoWriter(video_name, -1, fps, (width,height))
    for image in images:
        video.write(cv2.imread(os.path.join(image_folder, image)))
    cv2.destroyAllWindows()
    video.release()

# use opencv to read all images in a directory and compile them into a video
create_video('plot_directory', 'my_video_name.avi')

在create_video 函数中，我添加了反转帧顺序和设置每秒帧数 (fps) 的选项。 This video on Youtube 正是使用这种方法创建的。

要应用于您的示例代码，请尝试将所有绘图函数放入您的 for 循环中。这应该会生成您在边缘上迭代的每本书的情节。然后每次生成图时，您可以将该图保存到文件中。像这样的：

import random
from itertools import combinations
from math import sqrt
import itertools
from _collections import OrderedDict
import networkx as nx
import numpy as np
from matplotlib import pyplot as plt

random.seed(42)
n_points = 10


def dist(p1, p2):
    return sqrt((p1[0] - p2[0]) ** 2 + (p1[1] - p2[1]) ** 2)


points = [(random.random(), random.random()) for _ in range(n_points)]
named_points = {i: j for i, j in zip(itertools.count(), points)}

weighted_edges = dict()
tree_id = [None] * n_points
min_tree = []

for v1, v2 in combinations(named_points.values(), 2):
    d = dist(v1, v2)
    weighted_edges.update({d: ((list(named_points.keys())[list(named_points.values()).index(v1)]),
                               (list(named_points.keys())[list(named_points.values()).index(v2)]))
                           }
                          )

for i in range(n_points):
    tree_id[i] = i

sorted_edges = OrderedDict(sorted(weighted_edges.items(), key=lambda t: t[0]))
list_edges = sorted_edges.values()


for edge in list_edges:
    if tree_id[edge[0]] != tree_id[edge[1]]:
        min_tree.append(edge)

        old_id = tree_id[edge[0]]
        new_id = tree_id[edge[1]]

        for j in range(n_points):
            if tree_id[j] == old_id:
                tree_id[j] = new_id

        print(min_tree)


        G = nx.Graph()
        G.add_nodes_from(range(n_points))
        G.add_edges_from(list_edges)

        green_edges = min_tree



        G = nx.Graph()
        G.add_nodes_from(range(n_points))
        G.add_edges_from(list_edges)
        edge_colors = ['black' if not edge in green_edges else 'red' for edge in G.edges()]
        pos = nx.spiral_layout(G)

        G2 = nx.Graph()
        G2.add_nodes_from(range(n_points))
        G2.add_edges_from(min_tree)
        pos2 = nx.spiral_layout(G2)


        plt.figure(1)
        nx.draw(G, pos, node_size=700, edge_color=edge_colors, edge_cmap=plt.cm.Reds, with_labels = True)

        plt.figure(2)
        nx.draw(G2, pos2, node_size=700, edge_color='green', edge_cmap=plt.cm.Reds, with_labels = True)

        plt.show()