What controller? UC3842, or something else?
Looks very typical for their work. The ferrite beads are a dog's breakfast kind of thing, perhaps they actually needed all of them, or perhaps it's not worth optimizing further even given the quantities they're producing.
Which, let me see. Probably... fractional millions of units, for a given part number? Say another $10k in the lab to optimize EMI, so has to save say $0.05/unit x 200k qty to be worthwhile? They're probably getting those ferrite beads for pennies each (or fractions even?), and they're probably getting Chinese labor cheap enough to thread them onto the component leads for less than $0.05/ea, so, y'know, that's really not a terrible decision.
They do have enough of 'em, that replacing a bunch with one or two small (DM) chokes might be attractive. But maybe not quite enough so.
Shrug.
Anyway -- the design I would guess is 1-switch forward, hence the single primary side switch, and output choke; secondaries would be multi tapped, and rectifiers are per output. Possibly the TO-247 is for the 5V 6A output? The multi-winding choke filters all of them together (saves on cores, and helps a bit with cross-regulation innit?), plus either a charge pump (i.e. half wave voltage doubler) or a spare winding on the choke that generates -V, perhaps -15 to -18V, for a 7912, for the -12V output.
PCB is glass composite (FR-2 is it?), one layer, standard fare for cheap PSUs, and not so cheap as paper phenolic. Probably some EMI issues could be solved with better (2-sided) layout, but I would imagine they thought of that, and went with the cheaper option. (Perhaps threading ferrite beads really is just that cheap to do!)
Caps are brand name, so unless they've been counterfeited, they should be quite good, and MW has a positive reputation in the west so I would hope they are keeping track of things to maintain that.
I haven't been using anything long enough, that uses a MW power supply that I know of, to have even an anecdotal opinion on their reliability (which is perhaps a good thing, not having had a failure to speak of -- but it's also a small sample size, so not all that meaningful). As a designer, I certainly consider them acceptable.
Tim