Inspiration
As students who've always been extremely passionate about Computer science, and software development, we've found that it has become extremely easy to rely on AI to code for us. Looking around, this seems to be the case for most of our peers too! With this project, we aim to help students stray away from AI, while using AI. Because we need more good coders in the world, not prompt-generators.
What it does
DevPatch is a simple webpage that integrates a minimal IDE, and file management system. It also includes a messaging system with Bot, an AI whose goal is to help you learn. Bot is never allowed to truly give the answer away, but he can give tips, examples, and tricks while you're coding, to make sure you stay on track while developing your latest project. Bot analyzes your code, and actively gives you next-steps, and encouragement as you code. Bot actively recognizes errors you make while you code, and helps you fix them.
How we built it
We built the entire frontend using Javascript, HTML, and CSS. Bot is coded with Python, incorporating OpenAI's API for prompts, Watchdog to monitor file system events, and Pytest to easily create scalable test cases.
Challenges we ran into
Building the IDE with a working terminal was our biggest challenge, and fixing bugs with Bot brought a lot of stress!
Accomplishments that we're proud of
We are super proud of how well the DevPatch site turned out, especially for beginners to frontend coding! I am also proud of the logic behind Bot, and the use of WindowSockets to actively detect tokens while the user edits code. Bot's messaging system looks great!
What we learned
We learned that it becomes much easier to tackle backend problems, when the frontend is polished. Building the backend first led to many issues of creating the sockets and building a seamless connection between Bot, and the messaging system.
What's next for qhj - DevPatch
We plan on further developing DevPatch to add features such as a refined active code monitoring system. We plan on further improving the IDE, to incorporate extensions, more themes, more features, for an overall better user experience.
In the future, this project could scale to other areas of science and mathematics, to teach actively teach students other field, while using CV and Machine learning to detect hand-writing. Embedding into learning management systems (LMS) like Moodle, Blackboard, or Google Classroom. Adaptive learning models that analyze user progress and adjust explanations dynamically.
There are countless scaleable opportunities for this project, and given enough time, it will be something MASSIVE!
Log in or sign up for Devpost to join the conversation.