DETERMINING THE FLUX VALUE FOR SOURCES
For point sources, use 'imfit' to look for the centroid of the emission, in order to get the correct phase center (getting the position down to a couple of decimal points). As a general guideline, it is good to have a lambda/2 positional accuracy. Then use 'uvamp' or 'uvfit' with the 'offset' and 'center' keywords, respectively, in order to calculate the flux of the point source.
For extended asymmetric sources, the process is a little more complicated. You can use 'ellint' to sum up the flux, but this is best if the minor and major axes of the object are within a factor 2 of one another. With 'ellint', you want to plot the summed flux value for annuli of increasing radii until the cummulative flux flattens out.
For a more complicated source, sometimes consisting of an unresolved point source as well as extended emission that can be represented by a Gaussian, you can use 'uvfit' on a subset of baselines. First, determine the point source flux by fitting the longer baselines. Then, take the flux on the shortest baselines to be the total estimated flux. Subtracting the two numbers gives you the flux of the extended emission. You can also use a model to fit the shorter baselines, but this involves its own errors. And without single-dish (or zero-spacing) data, you cannot know the total flux exactly.