weather-forecasting

Project description:

A small WebAPI for weather forecasting and current weather information built with .NET 8, .NET Aspire, Redis. Due to the free Weather API limitations at the moment supports only 3-hour forecasts for 5 next days. The query rate limit is 60 per minute due to the same reason. A simple frontend web client is built with Angular.

Setup:

1. Clone the repository
2. Provide a OpenWeatherMap API key as an Environment Variable OpenWeatherMap__ApiKey or as a corresponding configuration in appsettings.Development.json.
3. Provide a "cache" connection string to a Redis server in the ConnectionStrings configuration section. Redis can be hosted in WSL2 in Windows, see https://redis.io/docs/latest/operate/oss_and_stack/install/install-redis/install-redis-on-windows/.
4. Run the project WeatherForecasting.AppHost via Visual Studio or via dotnet command line (dotnet run). At the moment there is no difference from running the WatherForecasting.WebApi project - Aspire just provides logs in a more convenient format. Later there will be Redis hosting and, hopefully, many other features.
5. Swagger documentation and interface are provided at {running_app_url}/swagger.

Running in Docker/Podman:

1. Publish the solution and build docker/podman images. (TODO: provide a Dockerfile)
2. Create a .env file with 'OWM_API_KEY' environment variable and paste your Open Weather Map API key as its value. The docker-compose process will use when the .env file and docker-compose.yaml are in the same directory.
3. Use docker compose or podman compose command to run the containerized application.

P.S. I used Aspir8 tool (https://github.com/prom3theu5/aspirational-manifests) to generate a docker-compose file from the Aspire project. This tool also builds images and publishes to your (local?) image repository so that the docker-compose command can actually run the container. It has some difficulties with secret management at the moment, so be aware to check if your secrets are populated properly.

Angular frontend application

Requires npm and node, I used npm v10.8.2 and Node v20.16.0. For client used angular packages with version 18.1.0.

Requirements and features:

Implemented

• The service should fetch current weather data from a public API (e.g., OpenWeatherMap, Weatherstack).
• Clients should be able to request weather data for a specific city/date.
• Use environment variables to keep the 3rd party weather API key.
• Implement the asynchronous operations.
• Implement exception handling with middleware. Respond with user-friendly meaningful error messages.
• Provide Swagger page.
• Provide automated tests for the service (unit / integration).
• Integrates logging (built-in logger), Application Insights (integration via Aspire, need to populated with data in WebAPI).
• Implement caching with Redis.
• Prepare the service to be containerized (Docker/Podman): added a Redis container to Aspire application model and referenced it in WeatherForecasting_WebApi. Now the application can run in a container (used Podman).
• Create a simple UI client using the Angular franework and include in the Aspire Host as a Node.js application. Currently supports only a limited set of cities. Mostly for my frontend leaning purposes.

Planned

• Implement error handling with retries.
• Enhance automated testing, provide more valuable test cases, implement integration tests.
• Decouple the forecast fetching service from the request handling service using a message queue (Azure Service Bus) to handle increased loads.
• Introduce authentication API (JWT) for the service. Maybe add some user management?
• Try to publish the API and the client in Azure. Deploy the weather forecasting service in Azure Cloud as AppService. Provide the script for Azure CLI for deployment or a Bicep file.
• Implement a CI/CD pipeline for the solution with Github Actions or Azure DevOps.

Under consideration

• Build something using Event Sourcing paradigm. For example: query 3rd party API (some scheduled operation), create a WeatherStateUpdate command and validate it, then publish CurrentWeatherStateUpdate event. Append the event to the log, in addition populate Redis cache or some document DB with the updated data. Think about some OLAP queries to run against the database with possibly a different model.