gp_treemap

Python wrapper around gp-treemap: turn a pandas DataFrame into a self-contained, interactive treemap HTML page. Mirrors the spirit of plotly.express.treemap for the path/values/color arguments, but emits the HTML directly — no Plotly Figure intermediate.

Install

pip install gp_treemap

The JS bundle and loader ship as package data, so the rendered HTML is fully offline: no CDN, no network fetch.

Usage

import pandas as pd
from gp_treemap import treemap

df = pd.DataFrame({
    "continent": ["EU", "EU", "AS", "AS"],
    "country":   ["FR", "DE", "CN", "JP"],
    "fuel":      ["nuclear", "coal", "coal", "nuclear"],
    "twh":       [380, 280, 5400, 70],
})

fig = treemap(df, path=["continent", "country", "fuel"], values="twh",
              color="fuel", title="2023 generation by fuel")
fig.write_html("out.html")   # standalone file
fig.show()                   # open in browser
fig                          # in Jupyter, renders inline (iframe srcdoc)

Arguments

• path: column names describing the hierarchy, outermost first.
• values: numeric column whose per-group sum drives cell size. Defaults to row-count.
• color: column name. Numeric → continuous palette; non-numeric → categorical hash. Pass "[Level 1]" (or leave color=None) to colour by the topmost visible ancestor — recomputed every time you zoom.
• theme, palette: gpdu themes / palettes ("tokyo-night", "nord", "viridis", "turbo", …).

Development

This package wraps the JS bundle built from the root of the repo. After running node tools/build.js at the repo root, refresh the bundled artifacts:

bash python/scripts/refresh_bundle.sh

Then run the tests:

cd python
python -m venv .venv && .venv/bin/pip install -e . pytest pandas numpy
.venv/bin/pytest