In db_init, when saving problems to db, parse_list_to_mathjson has been run so the tensor constants are already saved as mathjson in db. When problem is fetched from db, it tries to run parse_list_to_mathjson again.
-> In parse_list_to_mathjson of problem schema, if the first element is already 'List', parsing is not necessary anymore.
I have been using a similar bit of code in my local branch, which I assume both of us got from Matias. No issues with putting this into the main branch.
Problem:
-> In parse_list_to_mathjson of problem schema, if the first element is already 'List', parsing is not necessary anymore.