add a project manager so data for different experiments can be saved to project specific directories

Change the saved data format to something non-proprietary, possibly CSV?

migrate code base over to python 3, where all future features will be merged. Python 2 support should be dropped barring bugs and hotfixes

add a traditional eriksen flanker task

Add a new Settings option for defining a "Task quit" keyboard shortcut. All tasks can then reference this shortcut when quit functionality is needed

Refactor the sternberg code to allow users to manually set the number of trials per block. Currently it is hard coded, and not even passed into the class constructor

Each experiment should save data to its own directory. This would make it easier to run multiple experiments on the same computer

Data saving should be moved to the individual task modules, rather than all being done in the main script. This would make it easier to keep tasks modular

Related to #2 and #3

The starting image for Ravens matrices is incorrect as it loads all files in the images directory. However, this also includes a README, so all images are offset by 1

readme should be updated for MRT and ravens task images to demonstrate the correct method to extract partial images from existing image sources (e.g. purchased version of ravens)

supply some tools for aggregating data for each of the tasks

If task settings have already been saved, adding new task settings options in the future (for new tasks) will be a problematic

Default settings are only saved during "first run" (i.e. if the .ini file doesnt already exist). If a project is already in progress it will have settings saved (i.e. not first run). So adding any new task options in the future will not have defaults saved as the .ini already exists, so will error when trying to read from non-existing settings. The .ini cannot be deleted for existing projects because you will have to reset all the options again

May a solution is to try/catch the reading of potentially non-existing settings, and save defaults at that point instead of only during a "first run"

Certain tasks that load png images (e.g. MRT) show a warning about an incorrect sRGB profile:

libpng warning: iCCP: known incorrect sRGB profile

The fix is to convert all images to png without using ICC Profile: sRGB IEC61966-2.1

Some images will also warn about interlacing, so they should also be saved with no interlacing

The main BatteryWindow class should be moved into its own module

The current implementation checks whether a filename exists. However, this doesn't prevent someone from entering the same subject number multiple times with different condition numbers. Change so it checks for subject number instead of filename

add an n back task with controllable n

Size and position of main battery window does not save to settings.ini after a close event

make the tasks more modular so adding new tasks is easier and doesn't require code changes/additions in multiple places. Some sort of layer system would be nice for easy task contributions

add an options menu for each individual task where trials, blocks, and other options can be set and saved for future sessions