Text this: Sparsifying priors for Bayesian uncertainty quantification in model discovery