Incredibly well quantified emissions that are in total lower than the emissions from mining uranium (except for two or three cherry picked mines which are supposed to be representative), or the emissions from building and decomissioning a nuke if you take real lifetimes and load factors.
Most uranium ore is lower energy density than low grade coal. Digging it up with diesel equipment after removing twice as much overburden with explosives in a coal powered country and then milling it with 10s to 100s of litres of sulfuric acid is incredibly dirty. All of the "representative" lifecycle studies use Ranger (which used a specific much cleaner more expensive process only suitable for some specific ores on ore 30-70x as concentrated) or Cigar lake which is 1000-2500x as concentrated.
Even after that nuclear is still relatively low carbon, but about 10x a modern wind turbine. It is largely irrelevant (the best llw carbon technology is the one that deploys soonest), but that doesn't stop the shills constantly lying to try and delay decarbonisation.