main
Original Issue #29 was deleted by the maintainer. This fork preserves the evidence-based analysis showing the project's core CSI pipeline returns random data, models have no trained weights, and star count is suspected to be artificially inflated.