Building a Scientific Computing Platform (Quantum, ML, Math) in Pure Python Without NumPy

#algorithms #python #science #showdev

Hi everyone,

We rely so heavily on tools like NumPy, SciPy, and PyTorch that most of us treat them like black magic. I wanted to break that illusion. My goal was to
see if I could build a professional-grade computational science platform using nothing but the Python Standard Library.

It turned into an obsessive deep dive into algorithmic optimization, memory management, and raw math.

I've just released v0.6.0, and here is what it currently handles natively in Pure Python:

Linear Algebra & Calculus: I had to scrap naive 0(N!) determinants for an O(N3) Partial Pivoting LU decomposition to stop my machine from melting. Also
includes an Adaptive Newton Method.
Quantum Mechanics: A complete state-vector simulator. Instead of calculating matrices thousands of times for measurements, I implemented an 0(1)
probabilistic sampling algorithm with true state collapse.
Machine Learning from Scratch: A fully functioning Neural Network (MLP) trained by a custom-built Adam optimizer that features full momentum state
persistence.
Signal Processing: An O(N log N) zero-padded 1D/2D FFT and Convolution Theorem implementation.
Statistics: An autonomous hypothesis engine that takes generated ideas and tests them against datasets using ANOVA and Welch's t-tests.

The Ecosystem Upgrade:
To prove it wasn't just a toy, I made sure the engineering standards were flawless:

350+ unit tests hitting 95% coverage.
Strict typing enforced by mypy and ruff in the CI pipeline. Interactive Jupyter Notebook tutorials and a fully automated MkDocs website.
A Streamlit dashboard and a Rich-powered CLI.

If you are a student who wants to read the actual math behind an FFT or a Quantum Gate without getting lost in C/C++ source code, or if you're just
curious about how far vanilla Python can be pushed, I'd love for you to explore it. Feedback, code reviews, and stars are highly appreciated!

GitHub Repo: https://github.com/Furox88/cognitive-discovery-system

Documentation & Tutorials: https://furox88.github.io/cognitive-discovery-system/

PyPI Installation: pip install cognitive-discovery-platform