'Here b is an arbitary point on L, while a is the intersection of perpendicular line with q. By virtue of (5), $ <q\tilde{b}\tilde{a}=<qab= \frac{\pi}{2}$, $\tilde{b}$ lies on same line-segment $ q \tilde{a}$ as diameter. Done."
-Page 127 of Tristan Needham's book Visual Complex Analysis
note: (5) is property that triangle $ q \tilde{b} \tilde{a}$ and $ qba$ are similar.
I do not understand how the fact of two angles being 90 degrees and also triangles being similar implies that each point on the line has a unique map onto a point on the circle.


After about a year, I finally got what was going on here. I will write the steps of the proof pointwise.
We know that inversion is self inverse. So, if we prove that line gets mapped to circle, that's enough to say the circle goes to the line.
Suppose we have two fixed points $A$ and $B$, then the set locus of points $P$ for which $AP$ and $PB$ are perpendicular is a circle:
$\quad \quad$ Proof:
$\quad \quad$For simplicity, we can consider one of the fixed points as origin.
$ \quad\quad$ Given $A=(0,0)$ , $B=(B_x,B_y)$ , $P=(x,y)$, $AP \cdot BP =0$:
Join the line segment from center of inversion to nearest point of line $l_a$, call this point $a$. Now, choose an arbitrary point $b$ on the line $L$ and join a line segment $l_b$ from center of inversion to this point on the line $L$. Draw a perpendicular from $l_a$ to $l_b$. Also, let $ \tilde{a}$ on $l_a$ and $\tilde{b}$ and $l_b$ be the inverted points of $a$ and $b$ respectively.
Through properties of inversion, one can check that $ q \tilde{b} \tilde{a}$ is necessarily ninety degrees (refer). We also have the fact that $b$ was arbitrary, this means we have the case in mainpoint-2 meaning equivalently that we have a circle.
Therefore, circles get mapped to lines and lines get mapped to circles.