Normalization factor of joint density function

Question

I'm reading an article where authors define three density functions: $h_A(c)$, $g_A(y)$, $f_A(x)$.
Further in the text they define joint probability density function, given model $M = \{A_i\}$: $$P(x,y,c|M) = \frac{f(x)}{C(y)}\sum_{i=1}^ng_{A_i}(y)h_{A_i}(c)$$ where $C$ is a normalization factor: $$C=\sum_ig_{A_i}(y)$$ I understand what normalization factor is, but how did they get exactly this kind of normalization factor? What's the point of this formula of C?