我正在尝试使用 kivy Garden matplotlib 在 kivy 内创建测量可视化。目标是绘制字典中的变量及其相关测量值。
目前代码（见下文）由 3 部分组成。在导入部分之后，我定义了在代码中使用的常量。下一部分是一个字典，我用它以编程方式绘制具有不同设置（线条颜色、线条宽度等）的不同变量。从字典之后开始，我定义了一个方法切片数组来拼接数组以减少内存使用，并定义了一个类 Plots 来生成绘图。我还发了一个类似的问题 https://stackoverflow.com/questions/52200823/python-kivy-plots-with-meshlineplot-confusion但通过使用猕猴桃花园图。
我的代码有两个问题。首先，我不可能在没有的情况下显示坐标系。 ax.spines 命令没有任何效果。第二个问题是，我实际上希望能够绘制任意数量的变量，并将其显示在单独的子图中。我希望图形在显示屏上具有恒定的高度。这应该会导致滚动绘图的可能性。
import time
import psutil
import matplotlib.pyplot as plt
from kivy.app import App
from kivy.uix.boxlayout import BoxLayout
from kivy.clock import Clock
from kivy.uix.scrollview import ScrollView
from kivy.uix.widget import Widget
from kivy.garden.matplotlib.backend_kivyagg import FigureCanvasKivyAgg
# Chart Constants
MAX_MEASUREMENT_ARRAY_LENGTH = 500 # to trim measured values values
X_BASE_INTERVAL_LENGTH = 10 # time axis is displayed for an interval of 50 seconds
BASE_LINE_WIDTH = 0.5 # line width of lines in plots
BASE_LINE_COLOR = "white"
COLUMNS_OF_PLOTS = 1 # TODO implement more general layout possibilities, at the moment only 1-column
chart_variables = {
"cpu_usage": {
"order_number": 1,
"title": "CPU Usage",
"color": BASE_LINE_COLOR,
"line_width": BASE_LINE_WIDTH,
"x_interval_length": X_BASE_INTERVAL_LENGTH,
"y_bottom": 0,
"y_top": 100,
"x_label": "Time in epochs",
"y_label": "CPU usage in %",
"x_values": [],
"y_values": [],
"anti_aliased": False,
},
"upc_usage": {
"order_number": 2,
"title": "UPC Usage",
"color": BASE_LINE_COLOR,
"line_width": BASE_LINE_WIDTH,
"x_interval_length": X_BASE_INTERVAL_LENGTH,
"y_bottom": 0,
"y_top": 100,
"x_label": "Time in epochs",
"y_label": "UPC usage in %",
"x_values": [],
"y_values": [],
"anti_aliased": True,
},
"test3": {
"order_number": 3,
"title": "test",
"color": BASE_LINE_COLOR,
"line_width": BASE_LINE_WIDTH,
"x_interval_length": X_BASE_INTERVAL_LENGTH,
"y_bottom": 0,
"y_top": 100,
"x_label": "Time in epochs",
"y_label": "UPC usage in %",
"x_values": [],
"y_values": [],
"anti_aliased": True,
},
"test4": {
"order_number": 4,
"title": "test",
"color": BASE_LINE_COLOR,
"line_width": BASE_LINE_WIDTH,
"x_interval_length": X_BASE_INTERVAL_LENGTH,
"y_bottom": 0,
"y_top": 100,
"x_label": "Time in epochs",
"y_label": "UPC usage in %",
"x_values": [],
"y_values": [],
"anti_aliased": True,
},
"test5": {
"order_number": 5,
"title": "test",
"color": BASE_LINE_COLOR,
"line_width": BASE_LINE_WIDTH,
"x_interval_length": X_BASE_INTERVAL_LENGTH,
"y_bottom": 0,
"y_top": 100,
"x_label": "Time in epochs",
"y_label": "UPC usage in %",
"x_values": [],
"y_values": [],
"anti_aliased": True,
},
"test6": {
"order_number": 6,
"title": "test",
"color": BASE_LINE_COLOR,
"line_width": BASE_LINE_WIDTH,
"x_interval_length": X_BASE_INTERVAL_LENGTH,
"y_bottom": 0,
"y_top": 100,
"x_label": "Time in epochs",
"y_label": "UPC usage in %",
"x_values": [],
"y_values": [],
"anti_aliased": True,
},
}
def slice_arrays(max_array_length):
if len(chart_variables["cpu_usage"]["x_values"]) > max_array_length:
for key in chart_variables:
chart_variables[key]["x_values"] = chart_variables[key]["x_values"][int(max_array_length / 2):]
chart_variables[key]["y_values"] = chart_variables[key]["y_values"][int(max_array_length / 2):]
class Plots(Widget):
def __init__(self):
super(Plots, self).__init__()
self.number_of_plots = len(chart_variables)
self.t_start = time.time()
self.chart, self.axes = plt.subplots(nrows=self.number_of_plots,
ncols=COLUMNS_OF_PLOTS,
facecolor=(0.9, 0.9, 0.9))
plt.subplots_adjust(hspace=0.01)
self.chart.patch.set_alpha(0) # transparent plot background
def initialize_plots(self):
for ax, key in zip(self.axes.flat, chart_variables):
ax.set_title(chart_variables[key]["title"], color=chart_variables[key]["color"])
ax.tick_params(labelcolor=chart_variables[key]["color"])
ax.set_xlabel(chart_variables[key]["x_label"], color=chart_variables[key]["color"])
ax.set_ylabel(chart_variables[key]["y_label"], color=chart_variables[key]["color"])
ax.set_ylim(bottom=chart_variables[key]["y_bottom"],
top=chart_variables[key]["y_top"])
ax.set_frame_on(False) # remove white background inside each subplot
ax.spines['bottom'].set_color('white')
ax.spines['left'].set_color('white')
def plot_data(self):
slice_arrays(MAX_MEASUREMENT_ARRAY_LENGTH)
for ax, key in zip(self.axes.flat, chart_variables):
x_axis_interval_length = chart_variables[key]["x_interval_length"]
elapsed_time = time.time() - self.t_start
number_of_intervals = elapsed_time // x_axis_interval_length # index 0 to get div part only
chart_variables[key]["x_values"].append(elapsed_time)
chart_variables[key]["y_values"].append(psutil.cpu_percent())
ax.plot(chart_variables[key]["x_values"],
chart_variables[key]["y_values"],
color=chart_variables[key]["color"],
linewidth=chart_variables[key]["line_width"],
antialiased=chart_variables[key]["anti_aliased"])
self.chart.canvas.draw()
x_left = x_axis_interval_length * number_of_intervals
x_right = x_axis_interval_length * (number_of_intervals + 1)
ax.set_xlim(left=x_left, right=x_right)
def update_plots(self, _):
self.plot_data()
class MyApp(App):
def build(self):
plot = Plots()
plot.initialize_plots()
Clock.schedule_interval(plot.update_plots, 0.2)
scroll_view = ScrollView()
box = BoxLayout()
box.spacing = 24
scroll_view.add_widget(box)
box.add_widget(FigureCanvasKivyAgg(plt.gcf()))
return scroll_view
if __name__ == "__main__":
MyApp().run()
