Stan cannot load data as vector with an unknown size.
When using a python script, this is manageable as the size of the vector can be given to the processor.
However, when using a config file, things are hardcoded...
This is an issue relatively annoying as we will deal with data objects where the number of observations is unknown until execution.
A possible way is to define an hidden variable for each data which is a list: the pattern could be that, for data named myData, one could have an integer called _N_myData given to pystan.
Stan cannot load data as vector with an unknown size. When using a python script, this is manageable as the size of the vector can be given to the processor. However, when using a config file, things are hardcoded... This is an issue relatively annoying as we will deal with data objects where the number of observations is unknown until execution.
A possible way is to define an hidden variable for each data which is a list: the pattern could be that, for data named
myData
, one could have an integer called_N_myData
given to pystan.