Reinforcement Learning--Explanation to Formula (5.2)

The book doesn’t explain the formula (5.2) clearly, and the second and third lines of the formula (5.2) in page 101 made me confused. So, here, I make it clear to be understood.
First,
$$\begin{gathered} q_{\pi }(s,{\pi }^{\prime }(s))=\sum _a{\pi }^{\prime }(a\mid s)q_{\pi }(s,a) \\ \because \text{for all }\pi (a\mid s),\text{there is }\pi (a\mid s)=\{\begin{array}{ll}1-ϵ+ϵ/|\mathcal{A}(s)|& \text{if }a=A^{\ast }\\ ϵ/|\mathcal{A}(s)|& \text{if }a≠A^{\ast }\end{array} \\ \begin{array}{rl}\therefore q_{\pi }(s,{\pi }^{\prime }(s))& =\sum _{a(a≠A^{\ast })}\frac{ϵ}{|\mathcal{A}(s)|}{q}_{\pi }(s,a)+(1-ϵ+\frac{ϵ}{|\mathcal{A}(s)|})q_{\pi }(s,a=A^{\ast })\\ & =\frac{ϵ}{|\mathcal{A}(s)|}\sum _{a(a≠A^{\ast })}{q}_{\pi }(s,a)+\frac{ϵ}{|\mathcal{A}(s)|}{q}_{\pi }(s,a=A^{\ast })+(1-ϵ)q_{\pi }(s,a=A^{\ast })\\ & =\frac{ϵ}{|\mathcal{A}(s)|}\sum _a{q}_{\pi }(s,a)+(1-ϵ)\underset{a}{\max }{q}_{\pi }(s,a)\text{this is the second line of formula (5.2)}\end{array} \end{gathered}$$
Consider value $x$, let
$$x=\sum _a[\pi (a\mid s)-\frac{ϵ}{|\mathcal{A}(s)|}]q_{\pi }(s,a)$$
When $a≠A^{\ast }$, $\pi (a\mid s)=ϵ/|\mathcal{A}(s)|$
$$\begin{array}{rl}\therefore x& =[\pi (a=A^{\ast }\mid s)-\frac{ϵ}{|\mathcal{A}(s)|}]q_{\pi }(s,a=A^{\ast })\\ & =[1-ϵ+\frac{ϵ}{|\mathcal{A}(s)|}-\frac{ϵ}{|\mathcal{A}(s)|}]q_{\pi }(s,a=A^{\ast })\\ & =(1-ϵ)q_{\pi }(s,a=A^{\ast })\\ & =(1-ϵ)\underset{a}{\max }{q}_{\pi }(s,a)\\ & \le \underset{a}{\max }{q}_{\pi }(s,a)\end{array}$$
Also
$$x=(1-ϵ)\sum _a\frac{\pi (a\mid s)-\frac{ϵ}{|\mathcal{A}(s)|}}{1-ϵ}{q}_{\pi }(s,a)$$
$$\begin{array}{rl}\therefore q_{\pi }(s,{\pi }^{\prime }(s))& =\frac{ϵ}{|\mathcal{A}(s)|}\sum _a{q}_{\pi }(s,a)+(1-ϵ)\underset{a}{\max }{q}_{\pi }(s,a)\\ & \ge \frac{ϵ}{|\mathcal{A}(s)|}\sum _a{q}_{\pi }(s,a)+(1-ϵ)\sum _a\frac{\pi (a\mid s)-\frac{ϵ}{|\mathcal{A}(s)|}}{1-ϵ}{q}_{\pi }(s,a)\end{array}$$
This is the third line of formula (5.2). It’s clear to be understood now.