📅 Tuesday, 22 April 2025, 09:30-17:00
📍 Universitäts- und Landesbibliothek Darmstadt, Magdalenenstraße 8, 64289 Darmstadt
Sprints are collaborative, hands-on coding sessions that bring together open-source project maintainers and contributors to tackle real issues, ship features, and grow communities. Whether you’re a seasoned maintainer or a newcomer looking to contribute to open-source, our sprint day is the perfect place to connect and code together!
Think of sprints as a one-day hackathon—but focused on real open-source projects. You’ll meet maintainers, get onboarded quickly, and start contributing with guidance and support. It’s a great chance to:
✅ No PyConDE ticket required
❗ But registration is mandatory
Please get a sprint ticket at the PyCon DE & PyData 2025 website.
To prevent bots from taking the tickets, a credit card and 1€ s required for the reservation.
09:30 Welcome & Project Intro
10:00 Coding & Working Phase I
13:00 Pizza & Networking
14:00 Coding & Working Phase II
17:00 End
Here’s the list of projects confirmed for the 2025 sprints:
Coco is a "fitness tracker" for social interactions. It collects conversational data (through a hardware pendant) and quantifies this data and lets users interact with it through a frontend. All open source and fully self hosted.
https://github.com/mitralabs/coco
"VirtualVerse is a three-dimensional web-based virtual meeting world. It has the familiar look of a 3D voxel world (similar to games like ""MineCraft"") and allows you to modify every detail of your environment, block-by-block. You can also think of it as a 3D version of WorkAdventure."
https://gitlab.com/virtualverse
CARE is a collaborative research platform that streamlines studies and data collection across workflows like peer review. It features a PDF annotation system for inline comments and a built-in editor for seamless collaboration and editing in real time. Please note that the project's implementation will be in JavaScript.
Old repo, development continues at UKP Lab
https://eiwa.ukp.informatik.tu-darmstadt.de/docs/
"optimagic is a Python package for numerical optimization. It is a unified interface to optimizers from SciPy, NlOpt and many other Python packages.
optimagic’s minimize function works just like SciPy’s, so you don’t have to adjust your code. You simply get more optimizers for free. On top you get powerful diagnostic tools, parallel numerical derivatives and more. "
https://github.com/optimagic-dev/optimagic
Pi_Optimal is a Python library that enables data scientists to apply Reinforcement Learning with intuitively with only 7 lines of code. It works on tabular data, performs well especially with limited data, and requires no prior experience in RL.
https://github.com/pi-optimal/pi-optimal
"The Research Data Management Organiser (RDMO) is a web application supporting research projects in the planning, implementation and administration of all research data management tasks. It provides a collaborative platform for research groups with guided interviews to support researchers in complying with funding agencies‘ requirements.
RDMO is an Open Source project that started its development in 2015. It is build upon the Django Framework and leverages the Django REST framework to provide its API. The frontend is currently being migrated to React from Angular JS. The overall data model is split between an ensemble of models for the different elements of a questionnaire and the corresponding answers given by the users.
Users can have different roles for content management and collaboration within projects. The software supports the single-database multi-tenant deployment architecture through the Django sites framework. Furthermore, the integration with external services is supported through a flexible plugin architecture, which supports e.g. obtaining interview choices from external APIs or custom export formats.
The community is backed by the RDMO e.V., a cooperation of over 20 research institutions across Germany. The State and University Library Darmstadt is co-developing the software.
A running demo of RDMO is available at https://rdmo.aip.de/ (registration required). The software is extensively documented and a local development setup can be configured with the help of https://rdmo.readthedocs.io/en/latest/development/setup.html. " https://github.com/rdmorganiser/rdmo
Hyvis is a python package for visualizing (or more specifically scanning) high-dimensional functional landscapes, such as those created by loss functions in classical and quantum machine learning, and helps users develop intuition for the dynamics of their models. The core functionality is scanning the landscapes. That is, evaluating the function on some grid of parameters. Defining the grid (or the subspace) is the difficult part.
https://github.com/JoSQUANTUM/hyvis
Reliable LLM Memory for AI Applications and AI Agents https://github.com/topoteretes/cognee
"SynthIOS is an open-source pipeline for generating synthetic data for topic specific reasoning using powerful LLMs. It works with open-source models from the IONOS AI Model Hub, including Llama 3.1-70B, 3.3-70B, and 3.1-405B, or any other OpenAI SDK compatible Endpoint.
The SynthIOS pipeline consists of multiple functional blocks:
https://github.com/embraceableAI/SynthIOS
We use our self-sufficient weather stations to collect current data, combine this with meteorological data and weather data from the DWD and want to use it to forecast hotspots and temperature sinks in cities. We currently want to train our model to recognize trees in cities with DOP20 images and determine their vitality index with Sentinel2 data.