This is the source repository of my personal website / blog found at https://marcelltoth.net.
If you are interested in samples of my work, I recommend you to check out the Articles microservice for backend and the public facing Web SPA for frontend and also to read the articles on my site as they are a really good insight to the problems I am solving and the tech I am using day to day.
A while ago I decided that I would like to have a central place where I can share stories about my coding journey.
Of course my first idea was to install a good ol' Wordpress and customize that one - like everyone else. But I'm the guy who enjoys wandering off the beaten path, so I thought "Hey, I'm a web developer, why not build one for myself?". Of course it takes a lot of time, and it "does not make sense" in its traditional meaning, but I enjoy the process, and even the end result may be a bit better than I could've with Wordpress. (After all, a traditional web app will never match the responsiveness of a well-written SPA, and if you know me you know I'm all in for SPEED.)
First of all, you should know that this is not just "my blog". It's also my favorite playground where I can experiment with the very latest technologies and techniques in a real environment, a chance that I don't usually get with for-profit projects. I got to learn about Microservices, Domain Driven Design, NoSQL, Docker, .NET and ASP.NET Core 3, Ocelot, Swagger, React Hooks, PWAs, the Cache API, just to name a few.
I did not end up actually using all of them, but the experience was invaluable nevertheless. There are multiple techniques/solutions that I started learning here and now I use in my commercial projects to greatly improve the product, my productivity or both.
Because of the above, the disclaimer is: The architecture described here is way overkill for the size of this project and I would never ever recommend it. Way overengineering a product isn't much better than underengineering it, always match the solution to the project requirements / outlook. To me, this is a playground, so I kept developing (parts of) it as if it were a medium to large scale enterprise project. (Other parts are intentionally neglected to save my resources. Enterprise projects are not developed by a team of 1 for a reason after all. :D)
The project follows a somewhat traditional Microservices Architecture.
-
Bounded context: This service is an OpenID Connect provider that manages users in the system (aka me and myself) and issues tokens for the client applications to call the other services on their behalf.
-
Technology: Custom ASP.NET Core web application using IdentityServer v4.
-
Completeness: My goal was here to experiment with different OpenID Connect configurations, authentication modes. I am using in memory data and the default UI. (Currently I'm the only user, so an identity database would be kind of... overkill.)
-
Bounded context: This API manages the core blogging domain. Knows and cares about Articles and Tags.
-
Technology: ASP.NET Core API using DDD and CQRS.
-
Completeness: Mostly complete and functional. Missing is logging, monitoring, sophisticated error handling, etc., and the test coverage could be improved.
-
Bounded context: This API manages a static file storage, uploading and downloading of files. (Used mostly to embed images into articles.)
-
Technology: Simple CRUD ASP.NET Core API.
-
Completeness: In complete working order but "just as much as required", serves no demonstration purpose.
The project follows a BFF pattern in regard of API gateways, which stands for Backends For Frontends here rather than Best Friends Forever. Sad.
This means that I have an API gateway / aggregator for both clients, one for the Web SPA frontend that you can see when visiting my site, one for the Admin UI that only I can access.
This project uses the wonderful Ocelot library to perform basic mappings.
While I was OK with basic mappings for the Admin UI, I wanted efficient, customized mapping logic for the public frontend to reduce roundtrips and latency as much as I can.
This is a custom written ASP.NET Core web API that uses the simple HttpClient to fetch data from the services, aggregate them and push them down to the frontend.
This allows me for example to query two endpoints in parallel, and return it as a single result step, rather than having React preform 2 network requests / 2 reduces / 2 updates.
As I said before, I have two clients, a simple admin where I can write articles and upload files, and one for you to read them.
Technology-wise the two are similar, but only the public facing Web SPA project is detailed. If you want to take a look at a representative sample of my frontend work, check out that one.
The stack is based on: CRA / Redux / TypeScript / SCSS Modules.