identification big fixes

I made fixes related to #1733 (closed). the mod file now runs. there is still a bug however. I noted that since 4.6, Jacobian of moments are wrongly computed. In the model from https://forum.dynare.org/t/mode-check-plots-with-flat-lines/16020/13 the observation error shocks are obviously identified. this is correctly returned in 4.5, but the output of Iskrev method in 4.6/4.7 (i.e. equivalent of 4.5), provides a singularity. @wmutschl could you pls look into this? thanks