Describir: Universal subgradient and proximal bundle methods for convex and strongly convex hybrid composite optimization