The board bringup documentation is a good place to start:
Board Bring Up Introduction | TouchGFX Documentation
In general, TouchGFX relies on a way to read assets and display them on a screen, and it is up to the user to supply the means to that. The source of assets would usually be flash. If the flash is Q/O/HSPI, it will usually be memory mapped, in which case it is just read directly. Alternatively there is an option to supply a data reader to read, for example, normal SPI flash or EMMC.
Drawing on the screen is usually done in a framebuffer or two. If the LTDC is used to send data to a parallel RGB or DSI screen, code is generated to handle the framebuffers automatically. For DSI video mode and parallel RGB, this is usually enough, but for DSI a little custom code may need to be added.
For screens with SPI, FMC or other interfaces, it is up to the user to tranfer the pixels. This will usually be done using DMA, so the system is free for other tasks, like rendering the upcoming frame. For low-bandwith displays with internal frame memory, it is also possible to use a partial framebuffer that just updates the parts of the screen that changes from frame to frame, and does not hold the entire framebuffer in memory.
I hope this clears up where to start, at least. Since we already supply a Board Setup for that board, you can take inspiration from there. I would advise not to spend too much time on your own on the DSI stuff, it's a pretty complicated protocol that is very hard to debug.
