Hi all,
I believe I am hitting a bug. With some parameter choice, continuation of my model works fine. With some other choice, then EQ-C cannot complete and produces an IndexError on line 219 of BifPoints.py. Running a debugger, line 218 in the same routine produces an empty list, and it is not clear to me whether I should not have a empty a priori or there is a missing error check in the case in which the list W0 is empty...
M
Last edit: Maurizio De Pitta' 2016-08-15
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi, if you email me your entire script I can see if I can understand what the issue is. Sounds like maybe there should be a catch on that condition with a more reasonable report to the user.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi Rob,
This was a while ago... but I believe that this is the infamous unresolved bug in the detection of BP points pointed out by Drew in this thread.
I must say actually, that since you raise the issue, this bug currently prevented me from using PyDSTool and go back to XPPAUT after my last post, because indeed after trying to tweak with the tolerance as suggested by Drew, and trying to fool around a bit with the code, I miserably did not succeed. And yet (XPP)AUT in the bad (of the interface) and the good of native AUTO code, instead wins on this case. In conclusion, this is really a much needed fix.
M
Last edit: Maurizio De Pitta' 2016-10-30
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Yes, my apologies for being absent on this one - other job commitments kept me away from PyCont development. I agree that this is an important bug to fix. I'll get on it and update the thread mentioned above. Sorry for the LONG delay!
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi all,
I believe I am hitting a bug. With some parameter choice, continuation of my model works fine. With some other choice, then
EQ-Ccannot complete and produces anIndexErroron line 219 ofBifPoints.py. Running a debugger, line 218 in the same routine produces an empty list, and it is not clear to me whether I should not have a empty a priori or there is a missing error check in the case in which the listW0is empty...M
Last edit: Maurizio De Pitta' 2016-08-15
Hi, if you email me your entire script I can see if I can understand what the issue is. Sounds like maybe there should be a catch on that condition with a more reasonable report to the user.
Hi Rob,
This was a while ago... but I believe that this is the infamous unresolved bug in the detection of BP points pointed out by Drew in this thread.
I must say actually, that since you raise the issue, this bug currently prevented me from using PyDSTool and go back to XPPAUT after my last post, because indeed after trying to tweak with the tolerance as suggested by Drew, and trying to fool around a bit with the code, I miserably did not succeed. And yet (XPP)AUT in the bad (of the interface) and the good of native AUTO code, instead wins on this case. In conclusion, this is really a much needed fix.
M
Last edit: Maurizio De Pitta' 2016-10-30
Yes, my apologies for being absent on this one - other job commitments kept me away from PyCont development. I agree that this is an important bug to fix. I'll get on it and update the thread mentioned above. Sorry for the LONG delay!