basicnn 2026.4.13.10.53

A modular basic neural network toolkit with dataset and filesystem helpers.

basicnn

basicnn is a modular Python package with three focused submodules:

• basicnn.nn: neural network related abstractions and utilities.
• basicnn.dataset: general dataset and file-reading utilities.
• basicnn.filesystem: filesystem and path handling helpers.

Install

Default install (includes all module dependencies):

pip install basicnn

You can also install with explicit module extras:

pip install "basicnn[nn]"
pip install "basicnn[dataset]"
pip install "basicnn[filesystem]"
pip install "basicnn[all]"

For local development:

pip install -e .

For development tools:

pip install -e ".[dev]"

Auto Sync Dependencies

Use pipreqs-based sync script to avoid manual dependency maintenance:

python scripts/sync_deps.py

Optional mode examples:

python scripts/sync_deps.py --mode no-pin
python scripts/sync_deps.py --mode gt
python scripts/sync_deps.py --mode compat

The script updates these fields in pyproject.toml automatically:

• project.dependencies
• project.optional-dependencies.nn
• project.optional-dependencies.dataset
• project.optional-dependencies.filesystem
• project.optional-dependencies.all

Build package

python -m build

Upload to PyPI

python -m twine upload dist/*

One-Command Publish

Run this script to automatically:

• set version to YYYY.MM.DD.HH.MM
• build package artifacts
• upload to configured PyPI repository

python scripts/publish.py

Publish to TestPyPI:

python scripts/publish.py --repository testpypi

You can copy .pypirc.example to ~/.pypirc and fill your token.