I think the GEOSEARCH stage being labeled Easy is a bit misleading.
If the intended solution is to decode every point in the zset and brute-force the distance check against the query center, then yes, the stage is fairly straightforward. But if the goal is to understand or approximate how Redis actually makes geo queries efficient by leveraging geocells and sorted sets, this stage is much more subtle than “easy” suggests.
I’d suggest either:
- keeping the stage as Easy, but explicitly saying that a brute-force implementation is acceptable and expected for this stage, or
- bumping the difficulty to at least Medium (preferably Hard) if the stage is meant to push people toward a more Redis-like spatial indexing approach. It is probably also helpful to give some hints to push people in the right direction, otherwise it might not be immediately clear how the encoding and backing data store (sorted sets) actually helps with GEOSEARCH.
FWIW that was possibly the most difficult task I’ve ever done on this platform, required lots of research and meticulous implementation. Though I must say that it was also incredibly instructive and enjoyable 