Correct diskbb and change the way we compute integrals to enable gradient computation with JAX
Adding the AREASCAL and BKGSCAL when including background in the forward model
Automatic reparametrisation for transformed distributions in prior (e.g. LogUniform). It is much easier for NUTS to sample a Uniform variable which is then exponentiated (10^U(0, 5)) compared to a LogUniform variable
📚 Documentation preview 📚: https://jaxspec--121.org.readthedocs.build/en/121/