That doesn't make any sense. Observation is alteration. You cannot see light without absorbing light.

Right but how do you measure the things around what you are trying to measure and get any data from it unless you expect them to also interact with the things you are measuring.

You have to have an interaction to measure even if you are measuring the outcome and steps away from the original interaction.

It's like measuring dark matter where the easiest way to prove it's existence was to wait and capture the decay of it but not the particle itself. But that means the particle was already gone when we got the measurements to prove it was there.

It's not a physical wave. A wavefunction describes how likely something is to have different values for one of its properties. For example, an electron might have a wavefunction describing how likely it is to be in different locations. By observing if it actually is in a certain location or not, you force the electron to decide where it is concretely, "collapsing" the probability function into one value (its newly decided location).

Measuring is a loaded misnomer. Interacting with a particle changes what the particle is doing. There is no such thing as nondestructive testing in quantum physics.

Measuring just happens to be something we do a lot which necessarily causes particle interactions.