🎫 Example Issues Solved by Sweep

For Sweep's resume, see here.

PyTorch Ignite (opens in a new tab)

PyTorch Ignite uses Sweep for documentation fixes.

• Fix warning in fast_neural_style example (opens in a new tab) → Fix warning in fast_neural_style example (opens in a new tab)
• Docs test issue with roc_auc (opens in a new tab) → Update ROC_AUC Test Case in Documentation (opens in a new tab)

Medplum (opens in a new tab)

Medplum uses Sweep for small refactors and bug fixes.

• Sweep: fix missing key prop in FormPage.tsx (opens in a new tab) → Add missing key prop in FormPage.tsx (opens in a new tab)
• Sweep: prefer nullish coalescing operator in packages/app/src/HomePage.tsx (opens in a new tab) → Use nullish coalescing operator in HomePage.tsx (opens in a new tab)
• Sweep: Add support for env var config settings in packages/server/src/config.ts (opens in a new tab) → Add Environment Variable Config Support (opens in a new tab)
• Sweep: Add missing sort functions (opens in a new tab) → Implement Alphabetical Sorting Using localeCompare (opens in a new tab)

Sweep (opens in a new tab)

We use Sweep internally to fix small issues such as fixing a path to support windows and bug fixes like Python’s calling of git pull failing. We also have been doing setup things like GitHub Actions and ticket templates.

• Sweep: Use os agnostic temp directory for windows (opens in a new tab) → Use OS Agnostic Temporary Directory (opens in a new tab)
• Handle git pull failure in sweepai/app/ui.py (opens in a new tab) → Sweep: In sweepai/app/ui.py try except the git pull, if it crashes delete the cache and reclone it (opens in a new tab)
• Sweep: In on_comment.py it's possible that the comment is on a closed PR. If this is the case then we should not fire an event. Add logic to on_comment to address this. Use pr state (opens in a new tab) → Prevent Event Firing on Closed PRs (opens in a new tab)

Sweep.dev (opens in a new tab)

We also use Sweep for our landing page. We use Netlify to check for build completion and previews of our landing page and do iterations against the Netlify previews. This makes modifying the landing page possible with non-frontend devs like Will.

• Sweep: All of the testimonials say JE, can we change them to the correct initials (opens in a new tab) → Correct Testimonial Initials (opens in a new tab)
• Sweep: Change the last testimonial (the one by Kunal) to a new message (opens in a new tab) → Update Testimonial Content (opens in a new tab)

Plug ‘n Plai (opens in a new tab)

Eduardo (AI @ Stanford) is building Plug ‘n Plai, a marketplace for OpenAI plugins. He’s mainly using Sweep for single-file refactors like moving a chunk of code out of the script and small features like capping the max number of plugins.

• make .prompt a method and add verbosity and colors #140 (opens in a new tab) → Make .prompt a method and add verbosity and colors #141 (opens in a new tab)
• Create an auxiliary function in Plugins to parse a <API></API> message and call plugin #122 (opens in a new tab) → Add auxiliary function to parse API calls from LLM and call plugins #124 (opens in a new tab)
• Only allow activating a maximum of 3 plugins #65 (opens in a new tab) → Only allow activating a maximum of 3 plugins #75 (opens in a new tab)
• Cleanup the documentation in embeddings. The formatting is incorrect #89 (opens in a new tab) → Cleanup documentation in embeddings.py #91 (opens in a new tab)

ChangePack (opens in a new tab)

• Sweep: Redesign OmniAuth links in the registration form #145 (opens in a new tab) → Redesign OmniAuth Links in Registration Form #146 (opens in a new tab)

To contribute more, feel free to ping Kevin and Will ****on Discord.