CogZen

A small, modern toolkit of data-wrangling helpers for cognitive-science workflows — a 2026 rewrite of the original cogzen package.

Importing cogzen is cheap: heavy libraries (pandas, scikit-learn, IPython) are imported lazily, only when the helper that needs them is first used.

Install

uv pip install cogzen          # or: pip install cogzen
uv pip install "cogzen[dev]"   # + pytest, ruff

Modules

All public submodules are named aux_*:

Quick start

import cogzen
from cogzen.aux_str import clean_str
from cogzen.aux_str.now import now

logZ = cogzen.Log0()
log0 = logZ.logger
log0.info(now())
log0.info(clean_str("  messy   text\n\n "))

Package data, the modern way

The bundled Emacs logo is an example of how data ships and is read — no __file__ path arithmetic, works even from inside a wheel/zip:

from cogzen import aux_data

png = aux_data.read_bytes(aux_data.EMACS_LOGO_PNG)   # bytes
svg = aux_data.read_text(aux_data.EMACS_LOGO_SVG)    # str

with aux_data.path(aux_data.EMACS_LOGO_SVG) as p:    # real Path when needed
    ...

Develop

uv sync --extra dev
uv run pytest
uv run ruff check .
uv run ruff format --check .

License

GPL-3.0-or-later © cogsys.io