Skip to content

GITHUB_SETUP.md

Latest commit

 

History

History
156 lines (114 loc) · 4.48 KB

GITHUB_SETUP.md

File metadata and controls

156 lines (114 loc) · 4.48 KB
Raw

GitHub Setup Guide for Research Assistant

This guide will help you set up GitHub integration with automated CI/CD pipeline for the Research Assistant project.

🚀 Quick Setup

1. Create GitHub Repository

  1. Go to https://github.com/new
  2. Repository name: research-assistant (or your preferred name)
  3. Set as Public or Private (your preference)
  4. Don't initialize with README (we already have one)
  5. Click "Create repository"

2. Add GitHub Remote

cd /home/ramdov/projects/research-assistant
git remote add github https://github.com/nellaivijay/research-assistant.git

3. Configure GitHub Secrets

You need to add your Hugging Face token as a GitHub secret for automated deployment:

  1. Go to your GitHub repository
  2. Navigate to Settings → Secrets and variables → Actions
  3. Click New repository secret
  4. Name: HF_TOKEN
  5. Value: Your Hugging Face access token
  6. Click Add secret

To get your Hugging Face token:

4. Push to GitHub

# Push current code to GitHub
git push github main

5. Verify GitHub Actions

  1. Go to your GitHub repository
  2. Click on Actions tab
  3. You should see the CI/CD pipeline running
  4. The workflow will:
    • Run syntax checks
    • Install dependencies
    • Run basic tests
    • Deploy to Hugging Face (if on main branch)

📋 GitHub Actions Workflow

The CI/CD pipeline (.github/workflows/ci-cd.yml) includes:

✅ Testing Stage (runs on all branches)

  • Syntax Check: Validates Python syntax
  • Dependency Installation: Installs all requirements
  • Basic Tests: Tests imports and model configuration
  • Security Check: Scans for hardcoded credentials

🚀 Deployment Stage (runs only on main branch)

  • Automated Deployment: Pushes to Hugging Face Spaces
  • Triggered: Only after successful tests
  • Condition: Only on push to main branch

🔧 Workflow Triggers

The GitHub Actions workflow runs on:

  • Push to main: Full CI/CD + deployment
  • Push to develop: CI/CD only (no deployment)
  • Pull requests to main: CI/CD only (no deployment)

📊 Workflow Benefits

✅ Advantages

  • Automated Testing: Catches errors before deployment
  • Continuous Integration: Tests every change
  • Automated Deployment: No manual deployment needed
  • Code Quality: Ensures code standards
  • Security: Scans for credentials
  • Backup: GitHub serves as code backup

🔄 Development Workflow

  1. Make changes locally
  2. Test locally with python -m py_compile app.py
  3. Commit changes: git commit -am "Your message"
  4. Push to GitHub: git push github main
  5. GitHub Actions automatically:
    • Runs tests
    • Deploys to Hugging Face if tests pass
  6. Monitor in GitHub Actions tab

🐛 Troubleshooting

GitHub Actions Failures

If tests fail:

  1. Check the Actions tab for error details
  2. Fix the issue locally
  3. Push the fix
  4. Actions will automatically retry

If deployment fails:

  1. Verify HF_TOKEN secret is set correctly
  2. Check token has "Write" permissions
  3. Verify Hugging Face Space URL in workflow file

Manual Deployment

If GitHub Actions fail, you can always deploy manually:

git push origin main

This pushes directly to Hugging Face (current setup).

📝 Best Practices

  1. Branch Protection: Enable branch protection on main
  2. Pull Requests: Use PRs for code review
  3. Test Locally: Always test before pushing
  4. Monitor Actions: Check Actions tab regularly
  5. Secret Security: Never commit secrets or tokens

🔐 Security Notes

  • HF_TOKEN: Stored as GitHub secret (never in code)
  • API Keys: Use environment variables, never hardcode
  • Data Files: .gitignore prevents committing data
  • Logs: .gitignore prevents committing log files

🎯 Next Steps

After setup:

  1. ✅ Push code to GitHub
  2. ✅ Verify GitHub Actions run successfully
  3. ✅ Check deployment to Hugging Face
  4. ✅ Test the live application
  5. ✅ Set up branch protection (optional)

📚 Additional Resources

Your Research Assistant is now set up with professional-grade CI/CD pipeline! 🎉