Matplotlib-FigureStream¶
A backend for serve Matplotlib animations as web streams.
Instalation¶
pip install figurestream
Bare minimum¶
By default, the stream serves on http://localhost:5000
# FigureStream replace any Figure object
from figurestream import FigureStream
import numpy as np
from datetime import datetime
# FigureStream can be used like any Figure object
stream = FigureStream()
sub = stream.add_subplot(111)
x = np.linspace(0, 3, 1000)
# Update animation loop
while True:
sub.clear() # clear the canvas
# ------------------------------------------------------------------------
# Any plot operation
sub.set_title('FigureStream')
sub.set_xlabel('Time [s]')
sub.set_ylabel('Amplitude')
sub.plot(x, np.sin(2 * np.pi * 2 * (x + datetime.now().timestamp())))
sub.plot(x, np.sin(2 * np.pi * 0.5 * (x + datetime.now().timestamp())))
# ------------------------------------------------------------------------
stream.feed() # push the frame into the server
For fast updates is recommended to use set_data, set_ydata and
set_xdata instead of clear and draw again in each loop, also
FigureStream can be implemented from a custom class.
# FigureStream replace any Figure object
from figurestream import FigureStream
import numpy as np
from datetime import datetime
class FastAnimation(FigureStream):
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
axis = self.add_subplot(111)
self.x = np.linspace(0, 3, 1000)
# ------------------------------------------------------------------------
# Single time plot configuration
axis.set_title('FigureStream')
axis.set_xlabel('Time [s]')
axis.set_ylabel('Amplitude')
axis.set_ylim(-1.2, 1.2)
axis.set_xlim(0, 3)
# Lines objects
self.line1, *_ = axis.plot(self.x, np.zeros(self.x.size))
self.line2, *_ = axis.plot(self.x, np.zeros(self.x.size))
# ------------------------------------------------------------------------
self.anim()
def anim(self):
# Update animation loop
while True:
# ------------------------------------------------------------------------
# Update only the data values is faster than update all the plot
self.line1.set_ydata(np.sin(2 * np.pi * 2 * (self.x + datetime.now().timestamp())))
self.line2.set_ydata(np.sin(2 * np.pi * 0.5 * (self.x + datetime.now().timestamp())))
# ------------------------------------------------------------------------
self.feed() # push the frame into the server
if __name__ == '__main__':
FastAnimation()
Set host, port and endpoint¶
If we want to serve the stream in a different place we can use the
parameters host, port and endpoint, for example:
FigureStream(host='0.0.0.0', port='5500', endpoint='figure.jpeg')
Now the stream will serve on http://localhost:5500/figure.jpeg and due
the
0.0.0.0 host is accesible for any device on network.By default
host is localhost, port is 5000 and
endopoint is empty.
