eandklahn / molmag_ac_gui Goto Github PK

Currently, I'm reading new types of headers by making the program conform to the type of header - I should probably change it, so that the header is reformatted into a standard form that is then always used by the program

Should be implemented according to
https://pubs.rsc.org/en/content/articlelanding/2019/cp/c9cp04301b#!divAbstract

Closing the guess dialog without accepting does not stop execution of the fitting procedure - this has to be fixed

If input file has more than one susceptibility dataset (different applied magnetic fields), it should be possible to get all of these plotted in the T vs. XT plotting window in data treatment.

Instead of having a file called "dialogs", we should have a folder called "dialogs" with each custom dialog getting its own file.

And the same for other general things. A folder called "exceptions", where each custom exception is given its own file.

Similarly, let's collect the individual tab widgets in separate files in their own folder "tabs"

"process_ac.py" might still be it's own file in the "molmag_ac_gui"-folder, but we should also figure out what to do here. It is intended as a file that can be imported on its own with functions that are useful for scripting.

"utility.py" could be split into the other files, based on where these functions are useful.

It should be possible to choose a set of fitting parameters manually when doing fitting of relaxation times.

Idea: when doing a fit, the parameters that were automatically chosen should be presented in a subclass of QDialog. The parameters can then be inspected and changed if need be, before they are passed on to fitting by QDialog.accept()

It halts the GUI.

ACGui-class is a mess, and this can be helped by splitting the two tabs into separate classes

Currently, when some parameters are not used in the fitting procedure in the Data analysis tab, the guess value will still be shown in the fitting window. They are not used in the calculation, since the use parameter is set to zero, but it might confuse the user. These unused values should be zero in the window.

Problem with inserting columns with the same name in a Pandas dataframe (see below, not full error message)

File "C:\Users\emilk\Desktop\tutorial\lib\site-packages\molmag_ac_gui\ac_gui.py", line 668, in make_diamag_correction_calculation
... ValueError: cannot insert Mp_m (emu/mol), already exists

Something with QThreads...

Need to look at the data loading function again

After fitting χ' and χ'' in data treatment, the plot of the fitted data only shows the given function in the data points that are measured (i.e. for freq vs. χ', χ' is only calculated for each measured frequency, but it should be in a linspace in the frequency range). This will gives bumbs around the original datapoints. The plots can be smoothed out by calculating the function in additional data points

Currently, only ppms data files can be opened in data treatment tab. It should be possible to also open ac data files from SQUID (MPMS3). The function should use the second line of data file to identify the file type. This functionality is also usefull when we will implement DC functionality

I'd like a new feature

Needs to be calculated correctly for all types of data files

Instead of having "raw_df", "used_T", "not_used_T" and so on be attributes of the GUI MainWindow itself, there should be a data structure, so that accessing these variables would be something like

self.data.raw_df
self.data.used_T

In the process, can we come up with a consistent naming scheme for variables in the data?

When fitting AC data, it should be possible to change the fit resolution (number of points the log-space for plotting).

Make a temperature axis in the top of the plotting window in data analysis. This will make it easier for the user to decide which temperature range to fit in

Elements of the v, Xp and Xpp can be NaNs, and lmfit does not handle this

Currently, the data loading procedure is looking at number of frequencies by looking at the length of the set() of all frequencies, and that is then used to separate the data into blocks of equal size.
This assumes that the data is ordered into neat blocks, so if that is not the case, the data is grouped in wrong temp subsets.

To work with DC-data, many users use PHI to fit D and E parameters. It would be nice to have an option to get files out in the PHI format, i.e. _sus.exp and _mag.exp files (depending on the input data given by the user). This requires some manipulation of the units of moment and susceptibility.

Important that the GUI shows DC-data and only checks the file for AC-syntax when that route is attempted.