Skip to content

README.md

Latest commit

 

History

History
105 lines (74 loc) · 3.04 KB

README.md

File metadata and controls

105 lines (74 loc) · 3.04 KB
Raw

pequod · Python package

The Pequod palette for Python: the full Log base scale, eight crew accents (light and dark variants), and matplotlib helpers.

The narrative, design rationale, and full accessibility analysis live in the repository README and on the project website.

Install

From PyPI:

pip install pequod

With the matplotlib helpers:

pip install "pequod[plot]"

From source (the Python package lives in the python/ subdirectory of the main repository):

pip install "git+https://github.com/tiagojct/pequod.git#subdirectory=python"

Quick start

from pequod import LOG, CREW_LIGHT, CREW_DARK, palette

LOG                       # 12-step Log base scale, dict[str, str]
CREW_LIGHT                # 8 accents, light variants
CREW_DARK                 # 8 accents, dark variants

palette("log")            # full Log scale, list[str] of hex codes
palette("crew", n=5)      # first five crew accents
palette("log-cool", n=100, kind="continuous")    # interpolated to 100

Six named palettes:

Name Length Kind Suggested use
log 12 sequential continuous gradients, heatmaps
log-warm 6 sequential warm side only (Log 50–400)
log-cool 6 sequential cool side only (Log 500–950)
crew 8 qualitative categories, groups (light variants)
crew-dark 8 qualitative categories on a dark theme
syntax 8 qualitative crew in syntax-role order

Matplotlib

Two surfaces, depending on how much ceremony you want.

Register all palettes once, then use by name:

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

pequod.register_cmaps()    # adds pequod_log, pequod_crew, etc.

x = np.outer(np.linspace(0, 1, 200), np.ones(50))
plt.imshow(x, cmap="pequod_log")
plt.show()

Build a colormap explicitly:

from pequod import to_cmap

cmap = to_cmap("log", kind="continuous")        # LinearSegmentedColormap
cmap = to_cmap("crew", kind="discrete")          # ListedColormap
cmap = to_cmap("log-cool", reverse=True)         # reversed

to_cmap and register_cmaps raise a clear ImportError if matplotlib isn't installed; everything else in the package (the data dicts, palette()) is pure Python with no dependencies.

Keeping tokens in sync

The canonical palette lives at ../pequod.json (one directory above this package). The data shipped with the Python package is generated from it:

cd python
python data-raw/generate_data.py

That rewrites src/pequod/_data.py so the Python values always match the JSON tokens.

Licence

MIT. The underlying palette tokens are also available under CC-BY-4.0 via ../pequod.json; see ../LICENSE-CC-BY-4.0.