cosmoplots

Routines to get a sane default configuration for production quality plots. Used by complex systems modelling group at UiT.

Installation

The package is published to PyPI and can be installed with

pip install cosmoplots

If you want the development version you must first clone the repo to your local machine, then install the project and its dependencies with poetry:

git clone https://github.com/uit-cosmo/cosmoplots.git
cd cosmoplots
poetry install

LaTeX and Latin Modern dependency

The default style uses LaTeX to render text and equations. The default font is Latin Modern.

text.usetex : True
font.family : "Latin Modern"

Installations of TeX Live should include Latin Modern, but it may be manually downloaded from here.

Usage

Set your rcparams before plotting in your code, for example:

import matplotlib.pyplot as plt
import cosmoplots
# If you only want the default style
plt.style.use(["cosmoplots.default"])

Muliple subfigures

To make a figure with multiple rows or columns, use cosmoplots.figure_multiple_rows_columns. By default, the labels are $\mathrm{(a)}$, $\mathrm{(b)}$, $\mathrm{(c)}$, ..., but they may be replaced using the labels argument.

import matplotlib.pyplot as plt
import cosmoplots
plt.style.use(["cosmoplots.default"])

import numpy as np

rows = 1
columns = 2

fig, ax = cosmoplots.figure_multiple_rows_columns(rows, columns)
a = np.linspace(-3,3,100)
for i in range(rows*columns):
    ax[i].set_xlabel("X Axis")
    ax[i].set_ylabel("Y Axis")
    ax[i].plot(i*a)
# plt.savefig("assets/multifig.png")
plt.show()

Position of the y-axis

By default, the x-coordinates of the y-axis labels are aligned with the subfigure labels, but the y-axis label position can be adjusted using

ax[i].yaxis.set_label_coords(x_coordinate, y_coordinate)

change_log_axis_base

import matplotlib.pyplot as plt
import cosmoplots
plt.style.use(["cosmoplots.default"])
import numpy as np

a = np.exp(np.linspace(-3, 1, 100))

# Plotting
fig = plt.figure()
ax1 = plt.gca()
ax1.set_xlabel("X Axis")
ax1.set_ylabel("Y Axis")
base = 2  # Default is 10, but 2 works equally well
# Do plotting ...
ax1.semilogx(a)
# It is recommended to call the change_log_axis_base function after doing all the
# plotting. By default, it will try to infer the scaling used for the axis and only
# adjust accordingly.
cosmoplots.change_log_axis_base(ax1, base=base)
# Plotting
fig = plt.figure()
ax2 = plt.gca()
ax2.set_xlabel("X Axis")
ax2.set_ylabel("Y Axis")
base = 2  # Default is 10, but 2 works equally well
cosmoplots.change_log_axis_base(ax2, "x", base=base)
# Do plotting ...
# If you use "plot", the change_log_axis_base can be called at the top (along with add_axes
# etc.), but using loglog, semilogx, semilogy will re-set, and the change_log_axis_base
# function must be called again.
ax2.plot(a)
plt.show()

matplotlib vs. cosmoplots defaults

import matplotlib.pyplot as plt
import cosmoplots
import numpy as np

def plot() -> None:
    a = np.exp(np.linspace(-3, 5, 100))
    fig = plt.figure()
    ax = fig.add_subplot()
    ax.set_xlabel("X Axis")
    ax.set_ylabel("Y Axis")
    ax.semilogy(a)

# Matplotlib ------------------------------------------------------------------------- #
with plt.style.context("default"):
    plot()
    # plt.savefig("assets/matplotlib.png")
    plt.show()

# Cosmoplots ------------------------------------------------------------------------- #
with plt.style.context("cosmoplots.default"):
    plot()
    # plt.savefig("assets/cosmoplots.png")
    plt.show()

matplotlib	cosmoplots

generate_hex_colors

This function generates the hex numbers for the colours extracted from a matplotlib colour map based on the number of points of interest. The colors change gradually from bright to dark or vice versa.

import matplotlib.pyplot as plt
import cosmoplots
plt.style.use(["cosmoplots.default"])


color_list = cosmoplots.generate_hex_colors(5, 'viridis', show_swatch=True, ascending=True)
# plt.savefig("./assets/hex_colors.png")

# Print color_list to retrieve the hex numbers
print(color_list) #['#fde725', '#5ec962', '#21918c', '#3b528b', '#440154']

fig = plt.figure()
ax = plt.gca()
for i, color in enumerate(color_list):
    ax.plot([1,2],[i,i+1], c = color)

# plt.savefig("./assets/hex_colors_example.png")
plt.show()

hex_colors.png	hex_colors_example.png

combine

Sometimes, plots might be related and better placed as subfigures in a larger figure. If combining the plots using the subfigure environment in latex or similar is not an option, this is easily done with imagemagick in a systematic way.

Caution

This uses imagemagick v7.

The Combine class within the concat module implements such procedures, and is also conveniently available from the combine function in cosmoplots.

An example is shown below. Also see the tests directory for more examples. A help method that prints the imagemagick commands that are used under the hood is also available.

import matplotlib.pyplot as plt
import cosmoplots
plt.style.use("cosmoplots.default")
import numpy as np


def plot(i) -> None:
    """Create a simple plot."""
    a = np.exp(np.linspace(-3, 5, 100))
    fig = plt.figure()
    ax = fig.add_subplot()
    ax.set_xlabel("X Axis")
    ax.set_ylabel("Y Axis")
    ax.semilogy(a)
    plt.savefig(f"./assets/{i}.png")
    plt.close(fig)

plot(1)
plot(2)
plot(3)
plot(4)
plot(5)
plot(6)
plot(7)
plot(8)
plot(9)
plot(10)
# See `magick -list font` for all available fonts.
figs = [f"./assets/{i}.png" for i in range(1, 11)]
cosmoplots.combine(*figs).using(
    font="JetBrainsMonoNL-NFM-Medium",
    fontsize=30,
    gravity="southeast",
    pos=(100, 200),
    color="green",
).in_grid(w=3, h=4).with_labels(  # Specifying labels is optional
    "one", "four", "three", "two", "eight", "six", "seven", "five", "nine", "ten"
).save("./assets/concat.png")

# Note that cosmoplots.combine() == cosmoplots.Combine().combine()
cosmoplots.combine().help()
# Or equivalently
cosmoplots.Combine().help()