mpl-brandpacker

Package your brand design into a matplotlib.pyplot drop-in replacement. Define titles, headers styling, colors, and axes styling as Python classes — mpl-brandpacker handles the patching so your_brand.pyplot.subplots() returns fully branded figures and axes.

Install

pip install mpl-brandpacker

Quick start

1. Define your brand

# my_brand/colors.py
import mpl_brandpacker as mbp

class Colors(mbp.ColorsBase):
    primary = "#1a5276"
    accent = "#e67e22"
    dark = "#2c3e50"
    muted = "slategray"       # named colors work too
    cycle_0 = "C0"            # matplotlib cycle colors
    highlight = "coral"       # CSS named colors

Colors.plot()  # → swatch grid for documentation

ColorsBase accepts any matplotlib color string (#RRGGBB, #RRGGBBAA, #RGB, named colors, C0–C9, tab:blue, etc.) and normalizes to #rrggbb at definition time.

# my_brand/sizes.py
from mpl_brandpacker.sizes import MM_TO_INCH

class Sizes(mbp.FigsizesBase):
    half = (88 * MM_TO_INCH, 76 * MM_TO_INCH)
    full = (181 * MM_TO_INCH, 76 * MM_TO_INCH)

class FontSizes(mbp.SizesBase):
    title = 10
    body = 8
    footer = 6.5

    _scalers = {"presentation": 2.0}

SizesBase.scaled() is thread-safe — concurrent threads and async tasks each get their own scaled values.

# my_brand/figure.py
from mpl_brandpacker import BrandFigure, brand_method

class MyFigure(BrandFigure):
    @brand_method
    def set_title(self, title, **kw):
        self.mpl.suptitle(title, fontsize=10, weight="bold", x=0.02, ha="left", **kw)

    @brand_method
    def set_sources(self, sources, **kw):
        self.text(0.02, 0.02, f"Source: {sources}", fontsize=6.5,
                  color="#888", transform=self.transFigure, **kw)

    @brand_method(overwrite="savefig")
    def _branded_save(self, *args, **kw):
        """Override savefig with branded defaults.

        Pylance still shows the original Figure.savefig docstring
        because the implementation is underscore-prefixed.
        """
        kw.setdefault("dpi", 300)
        kw.setdefault("bbox_inches", "tight")
        self.mpl.savefig(*args, **kw)

Use @brand_method(overwrite="name") to override a built-in method while keeping IDE autocompletion on the original.

# my_brand/axes.py
from mpl_brandpacker import BrandAxes, brand_method

class MyAxes(BrandAxes):
    @brand_method
    def set_xlabel(self, label, **kw):
        self.mpl.set_xlabel(label, fontsize=8, **kw)

    @brand_method
    def set_ylabel(self, label, **kw):
        self.mpl.set_ylabel(label, fontsize=8, rotation="horizontal", **kw)

def set_style(ax, **kw):
    ax.grid(True, alpha=0.2)
    for spine in ("top", "right"):
        ax.spines[spine].set_visible(False)

2. Configure

# my_brand/__init__.py
import mpl_brandpacker as mbp
from pathlib import Path

mbp.configure(
    figure_cls=MyFigure,
    axes_cls=MyAxes,
    style_fn=set_style,
    stylesheet=Path(__file__).parent,  # dir with .mplstyle + data/fonts/
    pandas=True,                        # also hook df.plot()
)

3. Re-export pyplot

# my_brand/pyplot.py
from mpl_brandpacker.pyplot import *  # noqa
from mpl_brandpacker.pyplot import gcf

# pyplot-level shortcuts for brand methods (like plt.title in matplotlib)
def title(title, **kw):     gcf().set_title(title, **kw)
def subtitle(sub, **kw):    gcf().set_subtitle(sub, **kw)
def sources(src, **kw):     gcf().set_sources(src, **kw)
def footnote(note, **kw):   gcf().set_footnote(note, **kw)

For IDE autocompletion of brand methods on fig and ax, add a pyplot.pyi stub file — see the template for a complete example.

4. Use it

import my_brand.pyplot as plt

fig, ax = plt.subplots()
ax.plot([1, 2, 3], [4, 5, 6])
ax.set_xlabel("Quarter")

plt.title("Revenue")
plt.sources("Bloomberg")
plt.show()

Everything is branded. Original matplotlib methods are accessible via .mpl:

fig.mpl.legend()       # original Figure.legend
ax.mpl.set_xlabel()    # original Axes.set_xlabel

5. Reset (notebooks)

When iterating on your brand in a notebook, use reset() to start fresh:

import mpl_brandpacker as mbp

mbp.reset()                          # clears all hooks, reverts pandas
mbp.configure(figure_cls=..., ...)   # re-configure

What it provides

Export	Purpose
configure()	Wire everything up — one call
reset()	Clear all hooks (for notebook iteration)
brand_method	Decorator to auto-register methods for patching
BrandFigure	Base class for figure methods (subclass of Figure)
BrandAxes	Base class for axes methods (subclass of Axes)
ColorsBase	Color enum — accepts any matplotlib color string, normalizes to hex
FigsizesBase	Validated (w,h) size enum with .plot()
SizesBase	Font sizes with thread-safe context-managed scaling
PrintableEnum	Generic enum base

Submodules (advanced)

Module	Purpose
mpl_brandpacker.pyplot	Branded drop-in for matplotlib.pyplot
mpl_brandpacker.pandas	use_for_pandas() for df.plot()
mpl_brandpacker.patcher	brand_method, patch_method(), MethodProxy
mpl_brandpacker.style	register_stylesheet()
mpl_brandpacker.sizes	MM_TO_INCH, POINTS_TO_INCH
mpl_brandpacker.utils	get_text_bbox(), separate_kwargs()

How it works

configure() builds two functions — make_fig(fig) and make_ax(ax) — that patch matplotlib objects using Python's descriptor protocol. When you import my_brand.pyplot as plt, three pyplot entry points are intercepted:

• figure() → patches the new figure + wraps its axes creation methods
• gcf() → patches the current figure
• gca() → patches the current axes

All other pyplot functions (subplots, show, savefig, etc.) work unchanged because they internally call figure() or gcf().

@brand_method decorator

@brand_method                         # patches fig.set_title
def set_title(self, title): ...

@brand_method(overwrite="savefig")    # patches fig.savefig with this method
def _branded_save(self, *a, **kw):    # underscore name → invisible to Pylance
    self.mpl.savefig(*a, dpi=300)     # .mpl accesses the original

The overwrite parameter lets you override built-in matplotlib methods while keeping Pylance's autocompletion on the original signature.

Scaffold a new brand

python -m mpl_brandpacker.create_brand acme_corp --author "Jane Doe"

Generates a complete brand package from the built-in template with colors, sizes, figure, axes, header/footer layout, and a pyplot.pyi stub for IDE support.

Template

See template/ for a complete working example.