KosukeMizuno / socket-plotter

socket-plotter

Instant plotter based on pyqtgraph via socket communication.

Installation

with pipx

1. Install CLI commands.
   • pipx install socket-plotter[pyside]
2. Install in python environments where you want to use this library
   • activate your environment
   • pip install socket-plotter

Usage

from socket_plotter import plot_lines, plot_image, plot_image_and_lines
import numpy as np

# for line plot
xdata = np.arange(100)
ydata = np.random.randn(100)
plot_lines(xdata, ydata)

# for image plot
img = np.random.randn(100, 100)
plot_image(img)

# for both, the image and each row of the image are displayed.
img = np.random.randn(100, 100)
plot_image_and_lines(img)

The above three functions check if a plotter process exists. If needed, a new plotter process will be launched before plotting. You can specify your python executable to launch plotter GUIs via an environment variable SOCKETPLOTTER_PYTHON_EXECUTABLE.

Launch via command line interface

You can launch them as executables. Note that in this case SOCKETPLOTTER_PYTHON_EXECUTABLE will be ignore. The python interpreter in the installed environment will be used.

# launch it in background
$ socket_image_plotter &
$ socket_line_plotter &

Addresses and ports

The default ports are 8765 for lineplot and 8766 for imageplot. The default address is 127.0.0.1. Other ports and address can be assigned as the following:

plot_lines(xdata, ydata, addr='<address to plotter>', port=7777)

screenshots

Change log

[0.1.4]

• refine docstrings
• add entrypoints via CLI

[0.1.3]

• headers should be json-formatted.
• implement json-based data transfer

[0.1.2]

• executable switch via SOCKETPLOTTER_PYTHON_EXECUTABLE

[0.1.1]

• removed PySide2 from install_requires to avoid automatic installation by package managers for compatibility for pip and anaconda environments

[0.1.0]

• released

TODOs

• QThreadの適切な終了方法がよくわからないので終了時にエラーメッセージが出る
• socketのやりとりがキモいので修正したい