Most people instead buy pre-hung doors.

Bad example against reparation i.m.o. since when the hinge break they don't buy a new door(whole), but swap the broken hinge(part) for a new one.
Good example in the case of DIY though, since the hassle may not be worth the time spent.

most components can't be so easily replaced.

Every module in my computer, mouse, keyboard, screen, or, i.d.k., lamp torch, can be easily replaced with a screwdriver.
Even phones could be made easy to open. If you have a counter-example in mind to « unless when the part is difficult to access, which doesn't seem to be an engineering necessity in most cases ? » written above, then i'm interested.

Even if you have access to spare parts, it takes a lot of time to repair something even as simple as a radio.

But opening it and swapping the spare part(, well, welding it back then,) took less than 5mns.
What took a long time was opening it without breaking anything since it was fragile, with parts glued together. Radios were more complex than nowadays.
And they didn't stop at swapping the spare part apparently, but ran a full diagnostic because other parts aged as well and, e.g., a shorted transistor could overheat a transformer.
To me, it seems like asking for an individual to repair his watch himself by getting a spare part, these are the kind of situations that should be done by pros. But then even if it takes many hours we're not talking about a 20€ product, so it's usually worth it to repair instead of buying a new one(, which is why people repaired them instead of buying new ones).
Other examples could include houses or cars, which are repaired because buying new ones wouldn't be worth it.

But the example of the radio still goes in my direction, because back then it was difficult to swap the spare parts and yet people still went through the trouble of repairing it.
How much more would it then be pertinent for objects that are thrown away while a pro could easily swap the spare part in 5mns(, or an individual do it h·er·im·self).
The problem i.m.o. is that there's no repair·wo·man and no spare parts at a cheap price with close warehouses, so it's not profitable/widespread.
You're saying that most objects couldn't have their parts easily swapped while i'm saying that most could(, at least we agree that some can&can't)

Unless you can automate the entire repair process, increased automation will make us more likely to throw things away.

I also disagree that more human labor would be required to swap a part than to build the whole product 🤷
I even think that less human labor would be required to swap a part than to build it.
Many humans are involved with the production process, from mining to selling.
And if you were thinking in terms of advanced robots taking human jobs, then they'd eventually be able to do simple reparation as well(, and more&more complex over time).
(Edit : and if we could have said that «this automaton will create 150 circuit boards of this particular model every model, while the previous generation only made 50/h», we can also say that «this automaton will create 150 different kind of circuit board, while the previous generation could only create 50», if that's part of the counter-argument)

And worse, automation makes it easier just to start from scratch.

Not easier to build the whole than the part

You can always take a broken device, throw it in a crucible with a mountain of other broken devices, and just melt the whole lot down.

Not really :)
But you could strip its parts yes, that’d be the most sensical option if you can't repair it.