swak 1.1.2

Swiss army knife for functional data-science projects.

swak

Swiss army knife for functional data-science projects.

Introduction

This package is a collection of small, modular, and composable building blocks implementing frequently occurring operations in typical data-science applications. In abstracting away boiler-plate code, it thus saves time and effort.

• Consolidate all the ways to configure your project (command-line arguments, environment variables, and config files) with the cli and io packages, respectively.
• Wrap the project config into a versatile jsonobject.
• Focus on writing small, configurable, modular, reusable, and testable building blocks. Then use the flow controls in funcflow to compose them into arbitrarily complex workflows, that are still easy to maintain and to expand.
• Use the provided pandas and polars building blocks to get you started.
• Quickly set up projects on Google BigQuery and Google Cloud as well as AWS object Storage, and efficiently download lots of data in parallel with the cloud sub-package.
• Build powerful neural-network architectures from the elements in pt and train your deep-learning models with early stopping and checkpointing. From feature embedding, over feature importance, to repeated residual blocks, a broad variety of options is available.
• And much more ...

Installation

• Create a new virtual environment running at least python 3.12.
• The easiest way of installing swak is from the python package index PyPI, where it is hosted. Simply type

  pip install swak
  

  or treat it like any other python package in your dependency management.
• If you need support for interacting with the Google Cloud Project, in particular Google BigQuery and Google Cloud Storage, install extra dependencies with:

  pip install swak[gcp]
  

• Likewise, If you need support for interacting with Amazon AWS, in particular with S3, install extra dependencies with:

  pip install swak[aws]
  

• In order to use the subpackage swak.pt, you need to have PyTorch installed. Because there is no way of knowing whether you want to run it on CPU only or also on GPU and, if so, which version of CUDA (or ROC) you have installed on your machine and how, it is not an explicit dependency of swak. You will have to install it yourself, e.g., following these instructions. If you are using pipenv for dependency management, you can also have a look at the Pipfile in the root of the swak repository and taylor it to your needs. Personally, I go

  pipenv sync --categories=cpu
  

  for a CPU-only installation of PyTorch and

  pipenv sync --categories=cuda
  

  if I want GPU support.

Usage

Try making a new repository using the swak-template as a, well, template.

Documentation

The API documentation to swak is hosted on GitHub Pages.