Software projects

I really enjoy writing code, whether it is for data analysis, visualization or to automate small tasks in my daily life. Here, you’ll find a section with some of these projects. Most of them are work in progress and I usually don’t have the time to polish them to a point where I would consider them “finished”. I just continue to develop features as I need them.

chirpdetector

#

is a Python package that leverages YOLO models to detect electrocommunication signals on spectrograms of terrabytes of recordings of electric fish. The model backend is implemented as a “plug and play” system, allowing for easy integration of new models. The main selling point of this package is the framework to efficiently parse and process large amounts of data.

A previous prototype of this package using a simple CNN classifier was described in this article:

The current version of the package is still in development, but you can check out the repository for more information:

gridtools

#

is a Python package that provides a data model to load, process and save multi-channel electrode grid data including metadata and derived features, such as the communication signals and position estimates of the fish. The data model is designed to be easily extendable and implemented using pydantic to ensure type safety. I am also working on extending simulation features to generate synthetic data for testing and training purposes.

reportable

#

is a Python package that provides a simple way to “extract” reports from larger data science repositories. Sometimes, I find myself in a situation where I have a report, such as a slide deck, that links to a few plots in a directory with many other plots. This package helps me to extract only the relevant media files and copy them to a new directory and update the links in the report. This makes it easier to share the report with others without sharing the entire repository.

py2md

#

is a Python package that converts Python files to Markdown files. Let me explain: I hate Jupyter Notebooks. I really prefer to just have plain text files to make version control, and file management easier. So I came up with the idea to write my articles in Python files using block comments for text and the rest as code. This package extracts the text from the comments and wraps the code in Markdown code blocks. This way, I can write my articles in a single file and convert them to Markdown. An additional benefit is that the file is executable and easy to run, making it a self-contained and efficient solution. This article explains the process in more detail and at the same time, is the actual output of the package after running it on itself:

Check out the repository for more information:

slicepy

#

creates streamlined lists of important code edits. During rapid prototyping phases, I find myself writing tons of TODO comments which I then forget about. This package extracts these comments along with some context from a code base and then queries the OpenAI-API to generate a summary of every TODO item. The result is a markdown file with a checklist of all TODO items and their context, placed in the root directory of the project.

audioviz

#

is a Python package that provides a simple way to visualize audio data. Running it on an audio file, it creates an animated video of the spectrogram and the audio waveform in sync with the audio track. This package is still in the early stages of development, but I plan to extend it to include more features as soon as I need it for presentations or blog posts.

dotfiles

#

are a collection of configuration files for my Linux system. I use them to set up new systems quickly and to keep my system configuration consistent across different machines. This way, I can easily switch between my work laptop and my home desktop without having to worry about different configurations. They are highly opinionated and tailored to my needs, but feel free to use them as a starting point for your own configuration.