Find open source projects actively accepting contributors. Search repositories, filter by program milestones, difficulty tags, or tech stack.
Use our Orbit AI Matcher to find out! Get instant matching scores based on your developer skills, preferred frameworks, and contribution experience.
Convert your selected open-source project into a winning GSoC, LFX, or Outreachy application using Proposal Studio.
MetaSSR is a Server-Side Rendering (SSR) framework built entirely from the ground up during the Google Summer of Code (GSoC) 2024 program. It is a real-world use case for the MetaCall polyglot runtime, demonstrating how multiple programming languages can be utilized within a single SSR framework. Built with Rust and the Axum.rs web framework, MetaSSR is designed to offer developers a high-performance, versatile tool for creating SSR web applications.
This project aims to develop a Cables.gl plugin that compiles Faust DSP code into a WASM AudioWorklet in real-time. The plugin’s features will be focused on composability, modular I/O, and integration with Cables.gl’s WebGL backend. This is in addition to foundational development goals such as creating a pleasant UX, robustness, and elegance. This plugin would allow users to create DSP programs without boilerplate and have those programs integrate on a signal flow level with Cables.gl WebGL functionality. There are currently no tools that allow for this to my knowledge, this plugin would offer a unique opportunity for creative coding in both the audio and visual realms including simultaneously live-coding audio and visuals on one platform. This signal flow level integration would allow for passing arbitrary data to both the WebGL and WebAudio portions of Cables.gl programs, as well as between these different portions. This includes using shader buffers as audio input, enabling GPGPU for massively parallel audio computations, or audio output as sources for visual effects such as vertex displacement, SDF distortion, and other audio-reactive visual effects. This is in addition to simpler modulation schemes such as passing MIDI or OSC to both audio and graphics portions of programs in parallel or doing the same with Javascript sequencing.
Polyglot, a web application revolutionizing data exploration, offers dynamic 3D visualizations. Imagine zooming and rotating through a complex dataset, where each data point is a unique element. Polyglot empowers researchers to delve deeper into intricate relationships hidden within their data. This project proposes leveraging Polyglot's strengths to unlock hidden connections in medicine. By creating custom medicine embeddings that capture the relationships between drugs based on their shared salts (beyond just active ingredients), we can explore a massive dataset of over 600,000 medicines. Visualizing these embeddings in Polyglot's 3D space could allow researchers to identify previously unknown connections between medicines based on their salt compositions, potentially leading to breakthroughs in drug discovery, pharmacology, and personalized medicine.
Problems faced while developing labelled data: 1. Annotations made by a single person can include their biases into the data and eventually into the model 2. There are subjective tasks at times, which might not have a clear answer associated with them 3. The size of the dataset might be huge enough for a few people to annotate To resolve this, a solution is to let multiple people annotate the data and then aggregate the annotations based on a consensus. Crowdsourcing has become an essential paradigm for efficiently labelling large datasets, especially in the era of data-driven ML solutions. However, the challenge arises in aggregating the labels from a diverse pool of annotators, each with varying reliability. This project aims to develop solutions to handle the above problems and integrate their solution into CVAT. It
This proposal aims to address the lack of support for histogram charts and other advanced chart types introduced in Microsoft Office 2016 within LibreOffice. Histograms are essential for data analysis, scientific research, and machine learning, providing valuable insights into data distribution. Currently, LibreOffice users rely on some clunky workarounds to create histograms, limiting productivity and compatibility with industry standards. The goal is to implement native histogram chart support in LibreOffice, eliminating the need for workarounds and enhancing the user experience. By leveraging the existing column chart implementation as a starting point, the project will gradually transform and extend the code-base to support histogram-specific functionality. This approach minimizes disruption to the stable code-base while introducing new capabilities.
App Inventor 2 allows users to extend its functionality by using third-party extensions. Extensions are similar to built-in components, they can do everything from showing UI elements on the screen to performing tasks in the background. However, unlike built-in components, extensions are unable to have a mock preview in App Inventor's designer. The goal of this project is to make it possible for extension developers to be able to create mock previews for their extensions using web technologies, while also making sure that the security and stability of App Inventor is not compromised because of untrusted third-party code.
Project aims to integrate rmw_zenoh middleware to MoveIt2 and benchmark its effects on MoveIt2 project in points of message latency, performance of large data like pointcloud and how affects performance in situation number of topic etc. is changed. To modify CI jobs in a way being able to make it testable and benchmarkable for each ros middleware is also one of the aims of this project.
Users often encounter calls, SMS messages, and links across various mobile platforms, including browsers, social media, and messaging apps. However, when doubts arise regarding the legitimacy of these communications, such as suspicions of phishing attacks or the presence of malicious links, users currently lack a dedicated platform for submitting and reviewing such content. This absence of a centralized platform leaves users vulnerable to potential threats, as they lack a reliable mechanism to report and verify suspicious communications. Consequently, there is a pressing need for the development of a dedicated platform where users can submit calls, SMS messages, and links for review by trusted authorities or community members. Such a platform would empower users to proactively address security concerns, enhance collective awareness of potential threats, and contribute to the creation of a safer digital environment for all mobile users. To address this pressing issue, we propose the development of a comprehensive mobile application equipped with automatic detection capabilities tailored to identify and mitigate spams and scams. In addition to its automated features, the application empowers users to manually contribute instances of suspicious content encountered on diverse platforms, fostering a community-driven approach to combating online fraud. Complementing the mobile app, a dedicated web application will cater to administrative users, providing them with the tools and resources needed to conduct thorough reviews of reported instances. Through a systematic review process, pertinent information about identified spams and scams will be meticulously documented and stored in a secure database, facilitating further analysis and informing future detection strategies.
I propose to contribute to the OpenClimateFix organization by implementing the GenCast model (https://arxiv.org/abs/2312.15796), with a focus on advancing weather forecasting accuracy. By integrating state-of-the-art techniques such as diffusion models, sparse transformers, and graph neural networks, this project aims to address the challenges of improving weather prediction and contributing to climate research. The key deliverables of this project include: - Dataset Integration: incorporating the ERA5 dataset from WeatherBench2 to make comparison between different models easier. - Graph Mesh and Noise Generation: constructing the graph mesh and generating isotropic noise. - Model Architecture Enhancement: modifying the GraphWeather architecture, particularly focusing on the integration of sparse transformers in the Processor. - Conditional Layer-Norm Implementation: replacing standard layer normalization with conditional layer normalization. - Loss Function Adaptation: adjusting the loss function to incorporate noise level-dependent weighting. - Sampling: integrating the second-order DPMSolver++2S sampler for efficient score-based forecasts generation. - Testing and Documentation: adding tests during every phase of the project.
Doodle is a Scala library for 2D graphics, animation, data visualization, and creative coding. It supports JVM-based applications, utilizing Java2D for rendering, and for web browsers through SVG support. The library operates on a principle of composition (i.e Retained mode), enabling users to craft visuals by combining various elements. Doodle supports differences between backends through its tagless final architecture. Doodle currently supports retained mode graphics. Retained mode graphics involve representing the image as a data structure, creating a hierarchy of elements where each element relates to others. While this facilitates expressing parts of the image in terms of others, it poses challenges with large numbers of elements, making alignment difficult and memory usage proportional to the number of elements. On the other hand, immediate mode graphics represent the image as a function for drawing it. This method consumes memory proportional to the algorithmic complexity (Kolmogorov complexity) of describing the image. In immediate mode, we can have a single drawing context and can be used for high volume graphics. This enhancement will allow for more interactive experiences, opening up new possibilities for creative expression and application development. Further on, we wish to add advantages of the immediate mode, while retaining the advantages of retained mode. Combining both the methods can be very crucial for the developer willing to involve both the modes in a single method.
vertx-http-proxy is a programmable reverse proxy based on Vert.x. The current implementation uses interceptors to modify requests and responses during the HTTP calls. However, it’s inconvenient for developers to create interceptors for each rule they want to apply. This project will mainly focus on improving user experience by providing a builder API for transforming requests and responses from HTTP calls and enabling configure proxies by configuration files. Besides, it will also implement other useful features, such as cache storage integration and fine-grained management.
With Sktime’s goal of unifying machine learning with time series, I intend to contribute to that mission with the project titled “Scaling backends, foundation models, PyTorch, darts and pytorch-forecasting with sktime.” Where I’ll mainly be working on the integration with deep learning backends PyTorch through third-party library pytorch-forecasting and time series anomaly detection library darts. With the growth of the usage of deep learning networks in solving time series-related problems, it is needed to integrate sktime with the powerful deep learning libraries, and with this, I am interested in contributing to the Sktime project in this particular subject.
The first half of the project which is writing benchmarks and improving the webpage for storing it. In order to achieve this goal I will be using airspeed velocity to write benchmarks that cover almost all parts of tardis codebase (will take unit tests as reference). For improving the website, I will store more than 5 commits and add the commit which caused highest regression among the range of commits. The second half of the project is improving tardis codebase. In order to do this I will look out for the results from scalene and memray profilers. I will use faster algorithms, C/C++ libraries or functions and tools like mpi4py, polar etc to improve the codebase in terms of space and time complexity when running on both multi and single core.
The objective of this project is to enhance the PEcAn Project by optimizing its modules for standalone use. Despite PEcAn's robust framework and its interconnected modules, there's a growing need to make these modules independently operable. This shift is essential to simplify module usage, testing, and development, making the system more accessible and efficient for users and contributors. The focus is on optimizing our modules for standalone use, enhancing their individual operability within PEcAn's interconnected framework with our top priority being to 're-loosen these couplings by revisiting the design and interface of PEcAn packages .'
We have an online IDE that we currently embed into our problem pages. However, the IDE is currently embedded but not fully integrated, the reason being that it is not written in Vue.js + typescript as the rest of the web platform. As such it's hard to debug, add new features and work around with. We would like to refactor online IDE into a Vue.js component that is fully integrated into the problem pages of omegaup.com.
This project will implement and benchmark the Proximity Forest Algorithm for Time Series Classification in aeon. With the ever-increasing data, the applications of time series classification are also increasing. Hence, we need classification algorithms that are both efficient and scalable. The Proximity Forest Algorithm is the current state-of-the-art distance-based classifier that creates an ensemble of decision trees, where the splits are based on the similarity between time series measured using various parameterised distance measures. Currently, a version of Proximity Forest which can match the performance of the original implementation has not been implemented in Python. This project aims to implement Proximity Forest in aeon for the classification of univariate time series datasets of equal length and make it accessible for a greater variety of users. The implementation will be benchmarked on the UCR archive to match the results of the original Java implementation in terms of run time and accuracy.
The project aims to integrate KinFin, a tool for analysing protein family clusters, into genome browsing environments like Ensembl and GenomeHubs. This involves refactoring the KinFin codebase for better web compatibility by migrating to Python 3, enabling JSON output generation, and developing a "KinFin as a Service" (KaaS) system for user data processing. The goal is to enhance accessibility, visualisation, and data exploration capabilities for large-scale phylogenetic and genomic research by integrating KinFin's innovative approaches into widely-used genome browsing platforms.
PyZombis is a dynamic community course designed to teach Python to Spanish-speaking individuals, inspired by the Brazilian MOOC, Python para Zumbis. Delivered entirely online, PyZombis eliminates the need for users to install Python on their local machines or manage server configurations. However, the current implementation relies on Brython for executing Python within a browser environment, leading to inherent limitations. Certain functionalities, non-native Python modules, and C-built Python modules pose challenges for interactive learning experiences. The proposed solution aims to enhance PyZombis's interactivity and usability by addressing these limitations. The primary focus will be on PyZombis and the RunestoneComponents repositories. Key objectives include: Introducing a PyScript component to facilitate advanced exercises, enhancing interactivity and skill development, Implementing a Tkinter wrapper to enable browser-based execution of exercises, removing dependency on local installations, Enhancing user engagement through the implementation of a new user interface, fostering an immersive learning environmen, Strengthening the educative platform by integrating end-to-end playwright tests, ensuring reliability, and addressing pending issues for continuous improvement, Implementing stretch goals and overhauling documentation to further enrich the learning experience and accessibility. By addressing these objectives, the proposal aims to elevate PyZombis into a comprehensive and interactive learning platform for Python enthusiasts within the Spanish-speaking community.
PJBDD is a Java-based, multi-threaded Binary Decision Diagram (BDD) library that facilitates concurrent computation and parallel operations. It is specifically designed to enable efficient manipulation of Boolean formulas. This project aims to introduce a new type of decision diagram that is expected to improve performance by incorporating more reduction rules than existing ones called CESRBDD( binary decision diagrams with complemented edges and edge-specified reductions).
This project aims to improve the user experience of gprMax, an open-source software for simulating electromagnetic wave propagation by modelling Ground Penetrating Radar (GPR) and electromagnetic wave propagation, through the development of an AI chatbot and assistant. The purpose of the AI chatbot is to answer and troubleshoot any questions that the gprMax users may have, while the AI assistant is capable of turning natural language into an input file in the required gprMax format. We will leverage existing pretrained LLM models and the LangChain framework, and fine-tune the LLMs using LoRA and RAG techniques, drawing on the existing gprMax documentation and years worth of discussion in the Google groups and the GitHub issue tracker for data. The ultimate goal is to deliver and deploy an AI chatbot and assistant to enhance the accessibility of gprMax, as well as save both the users and the development team valuable time by automating the troubleshooting process.
Workbench is a multi-language code sandbox for GNOME. It is prevalent in the GNOME Ecosystem and is helpful to both professional and amateur developers. It provides a sane environment where you can quickly iterate on code in a GNOME application without going through the usual slow recompile cycles. Typescript is a language developed by Microsoft that adds static type annotations to JavaScript. This language has been trendy in the JavaScript/Web Development industry, and recently, we have seen projects like ts-for-gir that allow TypeScript to be used to build GNOME applications. Adding TypeScript support to Workbench would help developers write more robust code instead of the usual fragile JavaScript and help GNOME app developers get the most out of the growing TypeScript ecosystem by using new and modern features and syntax like decorators, iterables, etc.
Traditionally if a user asks a question to an LLM about a package then, the response might be incorrect, or outdated. The project "Supercharging the Knowledge Base of AIHelpMe.jl" is about extending AIHelpMe’s knowledge base and build a RAG system to include the latest information about its documentation and to increase the accuracy of LLM’s response. Incorporating comprehensive, up-to-date Julia’s resources into the AI’s knowledge base will benefit Julia developers, programmers, AI and ML enthusiasts, or any programmer to get more accurate responses, clear doubts and solve bugs.
The power of LLMs has been witnessed to be helpful to generate several short code pieces to fulfil the user’s demand, and be injected into channel chats or GitHub repositories. In this project, we propose to develop a Rocket.Chat app that enables the AI programmer functionalities where users can utilize the tool to create a short piece of code in C/C++, Java, Javascript, Typescript or Python based on specification supplied by the user. This implementation includes modern open source LLMs (Mistral, CodeLlama, WizardCode, Llama2, Phi, and so on) for a simplified configuration. This functionality also allows users to ask for a new variation of the code or augment/fine-tune the system for a more precise code generation. To achieve this, we propose to design a more intuitive and interactive UX to simplify users’ interaction. Finally, we will explore to bridge the generated codes with sharing APIs, allowing them to be shared to external applications and added to GitHub as a pull request. The AI Programmer app will be published on Rocket.Chat Marketplace together with specified documentation.
Kdenlive is a cross platform, open source, video editing software. Currently, Kdenlive saves its subtitles in the SRT format, which only allows very basic styling features - that means only one style for the complete subtitle file. A more advanced file format like ASS (SubStation Alpha Subtitles) would allow much more powerful styling, for example attributing different styles to different speakers, proper styling (using bold, italic or colors on words or sentences only), etc. Currently, Kdenlive can already read basic ASS files (without styling support). I'll improve Kdenlive's support for ASS files. Expecting results include opening and saving an ASS subtitle and keeping the styling infos, implementing a widget allowing to create styles and attribute a style to a subtitle sentence, and adding some attributes like bold to selected parts of the text.