E-ink Display Server Project

Overview

This project implements a Django-based server to manage and serve content to Raspberry Pi Zero 2 W devices equipped with e-ink displays. The system supports content updates, device authentication, and remote control via button presses.

Architecture

The project follows a client-server architecture:

• Server: A Django application with a REST API built using Django REST Framework (DRF). It manages display states, pre-rendered images, job scheduling, and device authentication. (This repository)
• Client: Raspberry Pi Zero 2 W devices with e-ink displays that poll the server for updates and send button press events. (Link to this repo will follow)

Key Components

1. Networking & Discovery

   • UDP broadcast for device discovery: Devices send EINK_DISCOVER requests, and the server responds with a bootstrap URL.
   • (Optional) mDNS/zeroconf for discovery.

2. Authentication & Security

   • DRF TokenAuthentication for device authentication.
   • Each device has a unique User and Token.
   • Bootstrap endpoint to provision new devices.
   • HTTPS for all communication.

3. API Endpoints

   • GET /api/display/state: Returns the current display state (mode, version, image URL, refresh intervals). Supports If-None-Match for efficient updates.
   • GET /api/display/image?version=...: Streams the image file. Supports conditional GET with ETag.
   • POST /api/display/button: Handles button press events and sets temporary overrides.

4. Rendering & Files

   • Images are pre-rendered and served via FileResponse.
   • Atomic writes to ensure image consistency.

5. Scheduling & Daemon

   • A single Django management command runs three async tasks:
     • UDP discovery server.
     • Stats collector (using psutil).
     • Job scheduler/executor (using croniter).
   • Time utilities for accurate scheduling (using timezone.now() and timezone.localtime()).
   • Database hygiene (closing old connections and offloading ORM operations).

6. Jobs System

   • Job model to persist job definitions.
   • Execution model to track job execution status.
   • Cron-based job scheduling.
   • Single-runner policy to prevent concurrent job executions.
   • Logging to the database, associated with job executions.

7. Photos / Variant Selection

   • Weighted random selection of image variants using PostgreSQL's with a scoring algorithm.

8. IDs & Bootstrap Payload

   • Human-readable display IDs.
   • Bootstrap response includes the display ID, token, and current mode.

Testing

python manage.py job "classify" --params-json "{\"max_num_to_classify\":10}"
python manage.py job "generate_variants" --params-json "{\"max_amount\":5}"

python manage.py job "get_weather"
python manage.py job "get_calendar"

python manage.py job "generate_dashboard"