Bovi

Monorepo for the Bovi dairy analytics platform.

Documentation lactationcurve package

Structure

Packages (libraries)

• packages/bovi-core/ — ML framework: base classes, registries, config, utilities
• packages/models/lactationcurve/ — Classical lactation curve fitting (Wood, MilkBot, etc.)
• packages/models/lactation-autoencoder/ — TensorFlow autoencoder for milk production prediction
• packages/models/bovi-yolo/ — YOLO object detection for dairy applications
• packages/infrastructure/pulumi/ — Azure infrastructure as code

Apps (deployables)

• apps/backend/api/ — Central FastAPI gateway with SQLite persistence
• apps/backend/models/lactation-curves/ — Azure Function App for classical curve fitting
• apps/backend/models/lactation-autoencoder/ — Azure Function App for TF autoencoder
• apps/frontend/dashboard/ — Next.js visualization dashboard

Getting Started

just sync        # Install dependencies + set up pre-commit hooks
just test        # Run all tests
just dev         # Start all services (API, models, dashboard)

just sync automatically configures git to use the checked-in pre-commit hooks (.githooks/), which run linting, formatting, type checking, and secrets detection on every commit.

Optional: If you use direnv, hooks are set up automatically when you cd into the repo.

Requirements

• Python 3.12
• uv — Python package manager
  • macOS/Linux: curl -LsSf https://astral.sh/uv/install.sh | sh
  • Windows: powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"
• just — task runner (install options)
  • macOS: brew install just
  • Windows: winget install Casey.Just, or download from releases
• bun — dashboard only
  • macOS/Linux: curl -fsSL https://bun.sh/install | bash
  • Windows: powershell -c "irm bun.sh/install.ps1 | iex"