Hello,
In addition to the one reported in #355 , other issues can be encountered when doing for instance model derivations using the symfit framework.
Starting for instance with a simple model1 defined with the expressiony(x; a, b) = a*x + b
After one derivation, the derivative model2 is created with the related expression: y(x; a) = a but you can't evaluate it with the same original parameters a and b : model2(x, a=1, b=2) because the b parameter is missing. This is quite embarassing when working with complicated model and n-order model derivation ... (with no idea of which parameters will still be present in the final expression).
Another problem is when evaluating the last expression y(x; a) = a with x as an array, model2(x, a=1) will return a 'singleton' [a] and not [a, a, ...a] of same size as x.
This could be fix in the future ?
Thanks
Patrick