- Handle version patterns like "19.2.x" and "5.x.x"
- Fetch all versions and find latest matching the pattern
- Resolve version ranges to actual version numbers before checking age
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Problem: Need to ensure all npm packages are at least 2 weeks old before use
Solution:
- Created check-package-age.js script to verify package publish dates
- Added .npmrc to enforce exact version installation
- Created pin-old-versions.sh helper script
- Documented complete workflow in NPM-PACKAGE-AGE-POLICY.md
Usage:
node scripts/check-package-age.js # Verify all packages ≥ 2 weeks old
npm ci # Install exact versions from lock file
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Removed deprecated chart files from helm/ root directory
- Updated all Helm documentation to reference warehouse13 chart
- Changed database name from 'datalake' to 'warehouse13' in values.yaml
- Updated helm command examples in SUMMARY.md
- Fixed migration instructions in helm/README.md
- Updated PostgreSQL backup/restore commands with correct database name
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Changed .Values.api.env.databaseUrl to .Values.app.env.databaseUrl
- This aligns with the unified architecture where api and frontend are combined into a single app
- Chart now passes helm lint successfully
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Changes:
- Replaced separate api and frontend deployments with single unified app deployment
- Updated values.yaml: Changed from api/frontend configs to single app config
- Renamed templates: api-deployment.yaml → app-deployment.yaml
- Removed frontend-deployment.yaml and frontend-service.yaml (no longer needed)
- Updated app image to warehouse13/app (multi-stage Docker build)
- Combined resource allocations: 384Mi memory, 350m CPU (up from separate totals)
- Updated all example values files (dev, production, air-gapped)
- Updated NOTES.txt to reflect single service on port 8000
- Updated ingress to route all traffic to single app service
- Added ARCHITECTURE.md documenting the unified container approach
Architecture:
The application now uses a multi-stage Docker build:
1. Stage 1: Builds Angular frontend with Node
2. Stage 2: Python FastAPI backend that serves static frontend from /static
Benefits:
- Simplified deployment (1 container instead of 2)
- Reduced resource usage (no separate nginx)
- Easier scaling (1 deployment to manage)
- Consistent versioning (frontend/backend always match)
Access pattern:
- http://localhost:8000 → Angular frontend
- http://localhost:8000/api → FastAPI REST API
- http://localhost:8000/docs → API documentation
- http://localhost:8000/health → Health check
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
