You might know I am printing ABS for while (Hints …). Now I printed a technical part in ABS and I made the surface smoother in acetone vapour according the proven method described at all3dp.com for about 30 minutes and let it dry about 2 hours. The part fits tight over a shaft. To slide it over the shaft easily I heated both to 100C in the oven. A disaster took place: enormous bubbles came out the part!
Definitely took care of your layer lines. 
-S
Modern art! 
Seriously, you might have better luck fitting it to the shaft first and then smoothing it with the shaft inside, if your setup permits and the shaft won’t be affected by the acetone vapour.
If you’re looking for an interference fit that material might not be a good choice for that. Better results might be had by allowing for a transition fit with a grub screw, or a snap ring, or roll pin, to retain it. I’m not sure heating the plastic to expand it, as you would with metal, would be as effective as freezing the metal shaft, or allowing greater clearance.
I don’t have much experience in this, but the few times I’ve put a metal shaft in plastic I went off these tolerances: Fits, offsets, and 3D printing
I’ve done an interference fit (0.1mm fit) and reamed to final size by hand for a transition fit, then retained with a mechanical fastener for strength.
With your square shaft a .15 or .2mm diameter delta may have better results for you.
Thanks for the link to that interesting article. (Maybe repeat that test with holes and pegs for the Snapmaker.) According the mentioned definition I made an interference fit!
