12 Hours, One Hackathon, and a Lot of Learning - Our Hacktoberfest Experience

说明与用途

在 Hacktoberfest 竞赛期间，团队构建了 DevTrace，一款利用本地 Ollama/Qwen AI 模型、Octokit.js 和 CI 流水线从 GitHub 讨论中提取并展示技术决策的工具。

问题

Understanding why a particular technical decision was made in open‑source projects, including rejected alternatives and their reasoning.

结果

DevTrace successfully demonstrated the feasibility of extracting and presenting engineering decisions from GitHub discussions within a 12‑hour hackathon, showcasing AI inference, GitHub API integration, and CI practices.

工作流

• 1. Define problem and scope during hackathon kickoff.
• 2. Discuss and select architecture: focus on AI pipeline, backend, GitHub integration, then frontend.
• 3. Build AI pipeline using Ollama with a small Qwen model to process discussion data and produce structured results.
• 4. Integrate GitHub via Octokit.js and GraphQL API to fetch repository, issues, PRs, and nested discussion comments.
• 5. Connect backend, AI, and GitHub data to the frontend to present decisions, alternatives, and evidence.
• 6. Set up CI pipeline to automatically test backend and frontend changes.
• 7. Iterate on debugging, communication, and feature refinement under time constraints.

证据

We wanted it to do more than summarise conversations. It needed to identify actual decisions, understand which alternatives were rejected and extract the reasons behind those rejections.

来源

We also decided to run our model locally using Ollama instead of depending entirely on external AI APIs. We experimented with a smaller Qwen model so that inference could run on our available hardware.

来源

We used Octokit.js to communicate with GitHub through its official APIs.

来源

Building a CI pipeline completely on my own was a new experience for me, and it was very good.

来源