Mpldock

A matplotlib backend that makes figures dockable.

Motivation

Work with something like this: instead of that:

Features

• Trivial integration - it is enough to change a matplotlib backend
• Arrange figures into a convenient layout with drag and drop
• Integrate mainloop with ipython, jupyter, pycharm console (or run as a standalone application)
• Preserve a layout and user-modifier state (scale, grid, etc.) through consecutive runs

Installation

pip install mpldock

Usage

Example

The simplest example with layout persistence:

import matplotlib.pyplot as plt
from mpldock import persist_layout

plt.switch_backend('module://mpldock.backend')
persist_layout('1e2682b5-4408-42a6-ae97-3c11332a96fa')

plt.figure("some plot")
plt.plot([1, 5, 3])
plt.figure("another plot")
plt.plot([5, 0, 1])

plt.show()

Set a backend

Set a matplotlib backend to module://mpldock.backend. See this for more.

E.g.:

import matplotlib
matplotlib.use('module://mpldock.backend')  # must be done before importing pyplot
import matplotlib.pyplot as plt

Or (does not work with jupyter):

import matplotlib.pyplot as plt
plt.switch_backend('module://mpldock.backend')

Also using a global configuration (like matplotlibrc file or MPLBACKEND environment variable) works but not in jupyter or pycharm console, since they hack around many things.

Integrate mainloop (if desired)

In a standalone application it is enough to use typical:

plt.show()

at the end of script, which starts a mainloop and blocks until the application is finished.

In an ipython or jupyter a gui magic may be used:

%gui qt5

Layout persistence

In order to preserve a layout between application runs, additional line must be added:

from mpldock import persist_layout
persist_layout('my_super_unique_identifier')

The layout is saved after closing a window or when done manually from a menu (Layout/Save). The string identifier should be different for each application (scripts with the same identifier share the layout).

More

See examples for more.

Major dependencies

• Python >= 3.6
• PyQt5 (PySide, PyQt4 coming soon)
• Matplotlib