search

qkitgroup / qkit

Qkit framework
GNU General Public License v2.0
43 stars 46 forks source link

It would be nice to store a hint in each h5 file on which entries to use for quick fitting while performing measurements. #48

Open JBraumueller opened 6 years ago

JBraumueller commented 6 years ago

This information could be read out by qfit and other post-processing modules.

rotzinger commented 6 years ago

This needs more explanations, I think. Jochen please explain it in more details. Thanks.

JBraumueller commented 6 years ago

My idea was the following: While performing a measurement, it is typically clear which of the dataset is the most relevant one. For instance in a single-port reflection measurement this is the phase most of the time, in time domain measurements, the phase_average is typically the one that is used to fit the data (freuqency of oscillations, damping, ...). While the dataset to be fitted is currently specified in the subsequent fit command, it would be nice to have a default stored in the respective h5 file that states which dataset to fit by default. Of course this can be overwritten by the fit command. By this, all fitting during the regular measurement procedure could be performed by one standard command.

Am 04.07.2018 um 12:08 schrieb rotzinger:

This needs more explanations, I think. Jochen please explain it in more details. Thanks.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/qkitgroup/qkit/issues/48#issuecomment-402518389, or mute the thread https://github.com/notifications/unsubscribe-auth/ANolNEOeJ5lYX2-rne3x1N5bNw6slkimks5uDOhtgaJpZM4T9yS1.

rotzinger commented 6 years ago

I see, that can certainly be implemented. It may be a good idea to make it unified over different measuring types, transport, spectroscopy, timedomain and so on.

yschoen commented 6 years ago

I think this is problematic and will lead to additional info stored fixedly into the h5 file that might in many cases be wrong. (Compare the trouble with comments that are forgotten to be adapted and forever preserve wrong information in h5 files). I think this should stay in the fitting command outisde the file.