"Taken to an extreme"? Is she implying there is a level of association fallacy that is not absurd?
Actually, I wanna respond to this. There actually is a level of assigning guilt by association that, while not necessarily logically rigorous, is nevertheless prudent. For instance, if I read about a Senator spending a lot of time at dinner parties and playing golf with, say, a member of the KKK, I would probably form a negative opinion of the Senator, even if he emphatically denied any sort of racist beliefs (voting habits, however, might convince me, but for the sake of argument let's say that looking at his voting record doesn't suggest a particular trend). Maybe the KKK dude's just got a killer sense of humor, but I don't feel that's a sufficiently plausible explanation to earn the benefit of the doubt.
While I would believe that worthy of suspicion, it's far from enough to make any sort of sound conclusion. I myself have friends with vastly different political beliefs than mine, and anyone who assumed I had the same opinions as my friends would just be laughed at.
As your example emphatically denies the things you're worried about, then I wouldn't think there's more reason to believe he's lying than any other politician.
Aren't there plenty of real life politicians who are friends or even married to people on the other side of the aisle? A random anecdote I'm reading right now says Bill Clinton and George H.W. Bush are buddies.