You are talking about detecting planets as they revolve around the star. As you said, if the planet rotates on an axis that "blocks" a tiny, tiny fraction of the star as we observe it, we see the dimming (that goes away as the planet passes) and can assume a planet and some parameters.
You are also right that the conversation is about COMMUNICATION: 1. You either dim the star with a Dyson swarm, or 2. You pulse more power than the star outputs as background noise briefly
Neither is all that viable.