Section 3.5 of "Kevin Patrick Murphy. Machine Learning: A Probabilistic Perspective." uses
$p(y_i|\pi)$ to denote the prior distribution in Bayes rule.
I've seen this kind of notation $Pr(y_i)$ a lot, to denote a discrete probability distribution, and rarely saw $p(y_i|\pi)$.
So, is $p(y_i|\pi)$ a common notation of discrete probability distribution