AttributeError>1traj=DS.compute('start_pts',dirn='c')/usr/local/pydstool/PyDSTool/Generator/Vode_ODEsystem.pycincompute(self,trajname,dirn,ics)130indepdom0,indepdom1=self.indepvariable.depdomain.get()131ifcontinue_integ:>132ifindepdom0>self._solver.t:133print("Previous end time is %f"%self._solver.t)134raiseValueError("Start time not correctly updated for "AttributeError:'ode'objecthasnoattribute't'
Bug or am I doing something wrong?
Thanks for your feedback.
Cheers,
M
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
The intention of 'continue' was not to allow this between python sessions. There is low level state in the DLL, _solver, that is not stored between sessions when you save the DS object, for instance the t attribute of the DLL. It's possible that I could get all of that transient state restored, but I'm not sure that all of it is accessible to the python layer. You could create an issue on github and I'll look into it.
An easy solution to support this inter-session is to make a second trajectory that begins from the end of the first, and then append their pointsets and reform a new trajectory from those. The internal state of the solver is still not the same, as restarting will invoke fresh choices of step size, etc., but it should be adequate for most purposes.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Dear all,
upon loading a generator
DS
byloadObjects()
and try to continue the trajectory, I obtain get a suspicious error.The code is:
And I get the following error:
Bug or am I doing something wrong?
Thanks for your feedback.
Cheers,
M
The intention of 'continue' was not to allow this between python sessions. There is low level state in the DLL,
_solver
, that is not stored between sessions when you save theDS
object, for instance thet
attribute of the DLL. It's possible that I could get all of that transient state restored, but I'm not sure that all of it is accessible to the python layer. You could create an issue on github and I'll look into it.An easy solution to support this inter-session is to make a second trajectory that begins from the end of the first, and then append their pointsets and reform a new trajectory from those. The internal state of the solver is still not the same, as restarting will invoke fresh choices of step size, etc., but it should be adequate for most purposes.