Time-Dependent Constraints in gfpop

This project extends the gfpop R package to support time-dependent constraints. Instead of using a fixed constraint graph, the graph can change at each data point based on a user-defined rule function. This makes it possible to express models like LOPART and labeled up-down peak detection within a single framework. Deliverables include the R API extension, C++ implementation, oracle-validated LOPART model, a 6-rule up-down model with label support, and a user-facing vignette.