search

stanfordLINQS / SQcircuit

Superconducting quantum circuit analyzer
BSD 3-Clause "New" or "Revised" License
43 stars 7 forks source link

Incompatibility with qutip 5 #23

Closed juanjosegarciaripoll closed 1 month ago

juanjosegarciaripoll commented 5 months ago

The package expects some qutip.qobj module that seems missing. I have only managed to make sqcircuit work by fixing qutip to 4.7.6

conda create -n sqcircuit python matplotlib-base
pip install qutip=4.7.6
pip install sqcircuit
hodgestar commented 5 months ago

QuTiP developer here. In the short term while SQcircuit does not yet support QuTiP 5, changing its qutip dependency to specify qutip~=4.7 seems like a sensible quick fix.

If there are questions regarding porting to QuTiP 5, feel free to ask here, or on the QuTiP mailing list or GitHub issues depending on which seems easiest / most appropriate. We did a fair amount of work to ease migration before the release, and we're keen to help as much as we can.

taha1373 commented 1 month ago

Thank you, @juanjosegarciaripoll, for using SQcircuit and bringing this issue to our attention. I also want to extend my thanks to @hodgestar for providing a quick fix. I'm pleased to inform you that we've released Version 1.0.0 of SQcircuit, which addresses the QuTiP dependency issue. Our latest version now fully supports QuTiP 5.0.0 and above.

hodgestar commented 1 month ago

@taha1373 That's awesome! How did the migration to QuTiP 5 go?

sambonkov commented 1 month ago

@hodgestar Overall moving to QuTiP 5 was mostly smooth. Before we relied on the QuTiP .data being stored as a CSR matrix, so it took a little bit to get used to the new options for data format and make sure we dealt with that correctly (without calling .full() and making it dense).

taha1373 commented 1 month ago

@hodgestar Apologies for the delayed response. I was waiting for @sambonkov's input since he handled the migration.

hodgestar commented 1 month ago

Glad to hear things went mostly smoothly. Woot.