You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, the only supported export option from a completed DeerLab FitResult is to pickle. Whilst pickling works well when everything is contained inside a single Python enviroment or computer it is not as good when transfering data between computer and Python versions. I am often finding that DeerLab fits completed on a spectrometer computer are not readable by my laptop.
My proposal would be to expand the FitResult class to contain an export method, as an additional option alongside pickling.
Advantages:
This method would then export to some heirachial format that would be more easily shared and stored.
This would also be much better for data archiving as instead of just saving a plot, the associated fit meta-data can also be kept.
Most people in our community (for better or worse) use Matlab as their main data-processing and ploting method.
Disadvantages:
Methods are not saved
UQResult would not be saved (probably could still save the important info but would take more work)
Some potential data format options:
HDF5 (Widly used in science but requires an extra package/dependency)
Old-Matlab (Already included in the scipy package)
New-Matlab (This is just a HDF5 wrapper)
Numpy / .npz (Designed for multiple arrays, and only readable in Python)
What do other people think?
The text was updated successfully, but these errors were encountered:
Something human-readable like TOML would be desirable. Python as an official TOML package as of version 3.11. Matlab doesn't natively support it, but there is at least one implementation available.
Currently, the only supported export option from a completed DeerLab FitResult is to pickle. Whilst pickling works well when everything is contained inside a single Python enviroment or computer it is not as good when transfering data between computer and Python versions. I am often finding that DeerLab fits completed on a spectrometer computer are not readable by my laptop.
My proposal would be to expand the FitResult class to contain an export method, as an additional option alongside pickling.
Advantages:
Disadvantages:
Some potential data format options:
What do other people think?
The text was updated successfully, but these errors were encountered: