SmartWeather is a microservices-based weather data acquisition system. It is designed to collect, process, and display weather data without relying on any existing weather stations. Instead, it integrates its own custom-built acquisition modules, a robust backend, and an interactive web interface.
SmartWeather is built as a monorepo with multiple layers to handle data acquisition, processing, storage, and presentation. The application supports containerization using Docker Compose.Kubernetes orchestration will maybe appear in the future but not immediatly.
Key Features:
- Embedded Systems Support: C++ code runs on ESP32 microcontrollers for data acquisition.
- Microservices Architecture: Each layer is modular to enable scaling and features additions.
- Real-Time Data Processing: Kafka and MQTT enable seamless streaming and communication.
- Flexible Storage Solutions: MySQL clusters for structured data and Elasticsearch for document-oriented data.
- Interactive Visualization: Kibana, Kafka-ui and a web client for monitoring and administration.
SmartWeather is built to be versatile and if you want to deploy a more robust and resilient version you can use the clustered version. This will automatically set up a master/cluster for the sql db and also for elastic search.
Most of SmartWeather services adopts an n-tier architecture with the following layers, for C# apps you will find theses layers in the /libs folder:
- Entities - Definition for the base entities that will structure application logic.
- Repositories - Persistence layer interface to abstract from databases implementation.
- Services - Main entrypoint for app integration, provides main functionalities to again abstract from repositories handling.
Each layer represent a library, it also exist for C++ application but it is not currently delocalized in the /libs folder.
SmartWeather.Project is made to be ultra portable, regardless of the complexity the project embded. Basically you will only need docker to run it, and a single command. If it's not the case, don't hesitate to open an Issue.
- Docker and Docker Compose installed.
- .NET SDK and Node.js (for development).
# Clone the repository
git clone https://github.com/your-repo/SmartWeather.git
cd SmartWeather.Project
# Start the entire application
docker compose -f ./app.yml up -d- Web Client: http://localhost:3000
- Kafka UI: http://localhost:8080
- Kibana: http://localhost:5601
| Folder | Description |
|---|---|
/Acquisition |
Embedded software in C++ and mock station in C# for data collection. |
/Communication |
MQTT brokers (HiveMQ, Mosquitto) and Kafka clusters for message handling. |
/Application |
C# applications including HTTP APIs, WebSocket APIs, and data historian services. |
/Persistence |
Databases: MySQL clusters and Elasticsearch for data storage. |
/Presentation |
Visualization tools (Kibana, Kafka UI) and a web client built with React/Next.js. |
/libs |
Shared C# libraries for entities, repositories, and services. |
/3dModels |
stl files that contains all 3d printable assets to build the station |
- Programming Languages: C++ (embedded), C# (backend), TypeScript (frontend).
- Databases: MySQL (clustered), Elasticsearch.
- Messaging Systems: Kafka, MQTT.
- Containerization: Docker, Docker Compose.
- Frontend Framework: React.js (Next.js).
- Monitoring Tools: Kibana, Kafka UI.
- Orchestration: Kubernetes (planned).
- [lucas066001] - Project Lead
- [matthias-goupil] - Front-end developer
- [DuquesnoyPaul] - Front-end developer
- [Az-r-ow] - QA
SmartWeather is currently not made to be open-source nor public, see the License for details.
Explore each subdirectory for more detailed documentation:
Happy coding! 😊