HTTP protocol URL REACT and NODEJS compared to using http protocol URL as file system

Question

So my question is about the URL as filesystem against the new web approach of NODEJS+EXPRESS and REACT.

lets assume that we have in the root directory '/' the following files:

root '/' directory:
    index.html
    image_super.jpg
    images_secret '/images_secret' directory:
        image1.png
        image2.png
        image3.png
        image4.png   

so we can approach to the files in the filesystem in the following way:

www.localhost:3000/index.html -> will get us the the index.html html form (as plaint text) www.localhost:3000/image_super.jpg -> will get us the image_super.jpg in the root directory www.localhost:3000/images_secret/image3.png -> will get us the image3.png inside the directory images_secret

and so on... what i'm trying to point that it's acting just like a file system! we're sending an HTTP 1.1 '/' GET request or HTTP 1.1 '/index.html' GET or HTTP 1.1 '/images_secret/image3.png' GET and it will get us the content right out our server file system.

just to mention - due to this filesystem URL we have LFI vulnerability that includes path traversal (that we can /../ go back and get out from the root directory)

and my confusion is: in NODEJS+EXPRESS we have middleware that treats our URL request not as filesystem we're treating our URL requests as Route middleware - for example we want to get /users for the URL it will use the Route middleware such as app.route('/users', getUsers) or somthing. so my question is: why when we're using NODEJS+EXPRESS it's not acting just like a filesystem? how does it possible that we use the URL as filesystem in one hand, and make it in the other hand a virtual URL that is not getting reall files?

same for react, i know that react is rendering the DOM objects in one HTML file SPA, but the URL is "virtual" there too.

Thanks!

URL to be one thing and not 2 different things, filesystem and as virtual abstract requests...