agplested / ascam Goto Github PK

This happens with the idealization frame where the button that says "Drag lines to change parameters" is too narrow to display the full text and with the names of lists where e.g. when using ascam -t the name of the baseline corrected, gaussian filtered list named "BC_GFILTER1000_" is not fully displayed.

for analysis, it is convenient to do threshold crossing only on the good traces. Therefore, a new feature to select the good traces and apply idealisation only to those is useful

Currently it's not immediately clear which histogram corresponds to the episode and which is the aggregate of all data.

When one repeatedly selecting Idealization from the Analysis menu multiple idealization windows are drawn on top of one another, each with their own parameters. Each of them needs to closed individually.

CK-filter can not be applied to the traces anymore. It is possible to type in the parameters but clicking 'OK' has no effect.

When trying to start ASCAM on a macOS Monterey I had problems. I am using a conda environment set up following the instructions on the README. I tried both using pythonw and using the automatically created ascam command.

Description:
After running ascam the standard logging information is printed and a window named "python" starts (appears in the tray). No error messages are displayed and the code runs through. The window isn't responsive and doesn't open (only appaers in the tray).

Solution:
The problem is solved by adding os.environ['QT_MAC_WANTS_LAYER'] = '1' in ascam.py 🙂

After loading, filtering, baseline-correcting and idealising a file the idealisation freezes from time to time when you go through the episodes after clicking 'Demo'.
Usually one or more episodes contain the idealisation of the previous episode (screenshots). To get rid of it one can click back and forth through the episodes, i.e. from ep. 170 (wrongly idealised with idealisation of ep. 169) to ep. 169 back to 170 until the idealisation is correct.
It appears that the setting of the dead time is correlated to that issue since it was not possible so far to reproduce the bug when no dead time was set.

To reproduce the problem I used the following procedure:

Load file 180426000_selected.mat

Import: 40kHz SR, ms time scale, pA current trace, V Piezo command, mV holding potential

filter 1kHz Gaussian filter

set baseline to zero -> poly

Idealize: Amplitudes 0, -0.6, -1.24, -1.84, -2.45 and resolution 0.08 ms -> Demo

At episode 18 (screenshot 1) I saw the first time that the idealisation of episode 17 (screenshot 2) was mistakenly used for 18. When I jump back to 17 (screenshot 2) and then to 18 (screenshot 3) the idealisation is correct.

ASCAM should raise an error if the Key field it empty. Otherwise if this happens accidently episodes keep getting added to and removed from the list.

It does however align if one moves the plot or zooms in/out.

Edit: Moving/zooming the trace plot disaligns the plots, moving/zooming the histogram plot aligns them.

in the idealization, when new values are entered but is not pressed before clicked 'apply and finish' the most recent values are not used
i.e. threshold crossing is not applied

I want to export data from ASCAM into a .mat file after doing baseline correction and filtering. When I do this and load the file in MATLAB, only one variable time is loaded (1x4000 double). Somehow when I open the file in ASCAM I get the complete data, baseline corrected and filtered.

Not obvious how to reproduce but exporting to matlab once threw exception that filepath cannot be a tuple (in the line using .endswith)

It seems that using >= for the requirement version was too optimistic. Some dependencies break when they are updated.
The easiest way to fix this is probably to pin the requirements to specific versions.

After finding the threshold values the plots get stuck after clicking 'Apply'. Switching to another series and then back resolves the problem. In raw you will also see the last idealized trace you viewed.

I could get around this by running pip install axographio==0.3.1 separately (but not with the newest version 0.3.2`. Maybe another case of incompatible newer versions.