The derivation of Bellman equation for value of a policy

In book ‘Reinforcement Learning - An Introduction’, Chapter 3, the author gives out the Bellman equation for $v_{\pi }$ as equation (3.14), but without detailed derivation. That makes me feel confused and uncomfortable, so I try to derive the Bellman equation by myself. The details of derivation are gave out as below:
$$\begin{array}{rl}{v}_{\pi }(s)& ={\mathbb{E}}_{\pi }(G_t\mid S_t=s)\\ & ={\mathbb{E}}_{\pi }(R_{t+1}+\gamma \cdot G_{t+1}\mid S_t=s)\\ & ={\mathbb{E}}_{\pi }(R_{t+1}\mid S_t=s)+\gamma \cdot {\mathbb{E}}_{\pi }(G_{t+1}\mid S_t=s)\\ & =\sum _a[{\mathbb{E}}_{\pi }(R_{t+1}\mid S_t=s,A_t=a)\cdot Pr(A_t=a\mid S_t=s)\\ & +\gamma \cdot {\mathbb{E}}_{\pi }(G_{t+1}\mid S_t=s,A_t=a)\cdot Pr(A_t=a\mid S_t=s)]\\ & =\sum _{a}Pr(A_t=a\mid S_t=s)[{\mathbb{E}}_{\pi }(R_{t+1}\mid S_t=s,A_t=a)+\gamma \cdot {\mathbb{E}}_{\pi }(G_{t+1}\mid S_t=s,A_t=a)]\\ & =\sum _a\pi (a\mid s)[\sum _{r}r\cdot Pr(R_{t+1}=r\mid S_t=s,A_t=a)+\gamma \sum _{g}g\cdot Pr(G_{t+1}=g\mid S_t=s,A_t=a)]\\ & =\sum _a\pi (a\mid s)[\sum _r\sum _{s^{\prime }}r\cdot Pr(R_{t+1}=r,S_{t+1}=s^{\prime }\mid S_t=s,A_t=a)\\ & +\gamma \cdot \sum _{g}g\sum _r\sum _{s^{\prime }}Pr(G_{t+1}=g,R_{t+1}=r,S_{t+1}=s^{\prime }\mid S_t=s,A_t=a)]\\ & =\sum _a\pi (a\mid s)[\sum _r\sum _{s^{\prime }}r\cdot Pr(R_{t+1}=r,S_{t+1}=s^{\prime }\mid S_t=s,A_t=a)\\ & +\gamma \cdot \sum _{g}g\sum _r\sum _{s^{\prime }}\frac{Pr(G_{t+1}=g,R_{t+1}=r,S_{t+1}=s^{\prime },S_t=s,A_t=a)}{Pr(S_t=s,A_t=a)}]\\ & =\sum _a\pi (a\mid s)\{\sum _r\sum _{s^{\prime }}r\cdot Pr(R_{t+1}=r,S_{t+1}=s^{\prime }\mid S_t=s,A_t=a)\\ & +\gamma \cdot \sum _{g}g\sum _r\sum _{s^{\prime }}[Pr(G_{t+1}=g\mid R_{t+1}=r,S_{t+1}=s^{\prime },S_t=s,A_t=a)\\ & \cdot Pr(R_{t+1}=r,S_{t+1}=s^{\prime }\mid S_t=s,A_t=a)Pr(S_t=s,A_t=a)/Pr(S_t=s,A_t=a)\left]\right\}\\ & =\sum _a\pi (a\mid s)\{\sum _r\sum _{s^{\prime }}r\cdot Pr(R_{t+1}=r,S_{t+1}=s^{\prime }\mid S_t=s,A_t=a)\\ & +\gamma \cdot \sum _{g}g\sum _r\sum _{s^{\prime }}[Pr(G_{t+1}=g\mid R_{t+1}=r,S_{t+1}=s^{\prime },S_t=s,A_t=a)\\ & \cdot Pr(R_{t+1}=r,S_{t+1}=s^{\prime }\mid S_t=s,A_t=a)\left]\right\}\\ & =\sum _a\pi (a\mid s)\{\sum _r\sum _{s^{\prime }}Pr(R_{t+1}=r,S_{t+1}=s^{\prime }|S_t=s,A_t=a)\\ & \cdot [r+\gamma \sum _{g}g\cdot Pr(G_{t+1}=g|R_{t+1}=r,S_{t+1}=s^{\prime },S_t=s,A_t=a)\left]\right\}\end{array}$$
$\because$ In Markov Process, $G_{t+1}$ only relate to $S_{t+1}$, $S_t$ and $A_t$ give no contribution to $G_{t+1}$,
$\therefore Pr(G_{t+1}=g\mid R_{t=1}=r,S_{t+1}=s^{\prime },S_t=s,A_t=a)=Pr(G_{t+1}=g\mid S_{t+1}=s^{\prime })$
$$\begin{array}{rl}\therefore v_{\pi }(s)& =\sum _a\pi (a\mid s)\{\sum _r\sum _{s^{\prime }}Pr(R_{t+1}=r,S_{t+1}=s^{\prime }\mid S_t=s,A_t=a)\\ & \cdot [r+\gamma \sum _{g}g\cdot Pr(G_{t+1}=g\mid S_{t+1}=s^{\prime })]\}\\ & =\sum _a\pi (a\mid s)\{\sum _r\sum _{s^{\prime }}Pr(R_{t+1}=r,S_{t+1}=s^{\prime }\mid S_t=s,A_t=a)\\ & \cdot [r+\gamma {\mathbb{E}}_{\pi }(G_{t+1}\mid S_{t+1}=s^{\prime })]\}\\ & =\sum _a\pi (a\mid s)\{\sum _r\sum _{s^{\prime }}p(r,s^{\prime }\mid s,a)\cdot [r+\gamma v_{\pi }(s^{\prime })\left]\right\}\end{array}$$
That’s the Bellman equation for $v_{\pi }$. We get it.