Yeah, I'm sure it's frustrating as hell. One of those moments where it would really help if these ECUs could show us live sensor data!
Running worse over time is likely just the plugs starting to foul. All six fouling the same suggests it's not an injector issue. The new fuel pressure regulator not helping suggests the excess fuel isn't coming up the vacuum line from a bad reg. So either something is forcing more fuel through the injectors than the computer intends, or the computer's getting bad info that's causing it to inject more fuel than it should.
Check that the fuel return line isn't pinched or otherwise blocked. Maybe a mud wasp got in there while you had it torn apart? If you can't blow through the return line, then it doesn't matter if the regulator is working, because the excess fuel can't leave the rail. I would also check that the fuel supply and return are plumbed the right way around, though I don't think it would run at all if you had those crossed up.
If the fuel lines are OK, try running it with the coolant temp sensor disconnected. If the computer knows that the temp sensor isn't working, it'll guess the engine temperature based on how long it's been running. If this makes it run better than it did with the sensor hooked up, check the resistance of the sensor, and the wiring between the sensor and the computer. (The service manual's troubleshooting for the temp sensor code should tell you which pins to probe for this.) If that doesn't help, plug the temp sensor back in and try the same with the MAF. Again, if it runs better on guesses than it does on sensor data, check everything in that circuit. I know it stalled when you tried unplugging the MAF while it was running, but that may be different it you start it with no MAF vs if it loses connection while running.
Check that it codes for the temp sensor and MAF when they're unhooked. If the computer doesn't notice that a sensor is unplugged, then something ain't right with either the harness (current leakage makes the computer think it's seeing sensor data?) or the computer. I would also disconnect the battery for a bit before/after each test to clear any previous codes and self-learning.
I would not expect a blocked exhaust to cause a rich condition.