🐐 Goat Master

A comprehensive, hybrid ML-based goat herd management system built with Next.js, TensorFlow.js, and PostgreSQL (pgvector).

✨ Features

• Offline-First Architecture: View and edit data without internet access using IndexedDB local storage. Changes are queued and automatically synced to the server in the background when connectivity is restored.
• Comprehensive Herd Management: Dedicated modules for Profiles, Inventory, Sales (with PDF receipts), Lineage/Breeding, Health records, and automated Alerts.
• Smart ML Recognition:
  • On-device: TF.js (MobileNetV3 + custom re-ID) for instant local matching.
  • Server-side: FastAPI + YOLOv8 + ResNet50 for high-accuracy verification.
  • Vector Search: PostgreSQL pgvector with <=> cosine distance operator.
• Secure Authentication: Built-in JWT authentication with secure HTTP-only cookies and bcrypt password hashing.

🚀 Getting Started

This project uses Bun for high-performance dependency management and execution.

Prerequisites

• Bun v1.0+
• PostgreSQL with pgvector extension

Installation

bun install

Environment Setup

Create a .env.local file based on your credentials:

DATABASE_URL="postgres://user:pass@localhost:5432/goatmaster"
JWT_SECRET="your-secret-min-32-chars"
MIGRATE_SECRET="migration-key"
ADMIN_USERNAME="admin"

# Optional: ML Microservice
ML_SERVICE_URL="http://localhost:8000"
ML_SERVICE_KEY="your-ml-api-key"

# Optional: Cloudinary image uploads
NEXT_PUBLIC_CLOUDINARY_NAME="your-cloud-name"
NEXT_PUBLIC_CLOUDINARY_PRESET="your-preset"

Database Migration

Initialize the schema, tables, and enable pgvector:

curl -X POST -H "Content-Type: application/json" \
  -d '{"secret": "your-migration-key"}' \
  http://localhost:3000/api/db-migrate

Development

bun dev

🧠 ML Architecture Details

• Training: The re-ID model is fine-tuned on all enrolled goat embeddings (shared model across users) using triplet loss.