Hello all,
As it currently stands, pylinac has the modules I originally conceived of. There are other modules I am considering, but that is as far as they have gone. I originally was going to construct a roadmap, but instead I’d like to open up the floor to see what users, or potential users, would like to see in pylinac. I’m currently seeing two fronts where work can continue.
The first is more modules. Pretty straightforward, but obviously the more pylinac can do, the more one would be willing to use it. Pylinac is not a one-stop-shop for all TG-142 QA, but I’d like to see it progress in that direction.
Right now I can see a few modules that may prove useful: Flatness/Symmetry (either film or EPID w/ baseline), Imaging (e.g. kV/MV planar phantom analysis).
What other modules would be useful?
The second direction is that of platform. I’d be happy to keep pylinac as a python module that tinkering physicists can use, but more people would be willing to use it if it was more accessible. One option is to make a GUI wrapper. Non-programmers would clearly be drawn to such an option and it wouldn’t be that hard to implement. One drawback of sinking work into a GUI is the clear direction we are all facing: cloud-based QA and databases. Another option is to make it web-based. A website allowing a user to upload their images and performing automatic analysis sounds pretty sexy to me, although I don’t yet know Django or other web frameworks which would prolong the time from conception to execution. Another option is integration with existing technology. The only project I can think of that might currently provide shelter for pylinac on the web is QATrack+, an awesome project that may benefit by synchronizing certain QA tests and their analyses through pylinac.
So then, should pylinac branch out into its own or try to merge with other projects?
I realized yesterday that you opened the discussion here about some future direction.
At my clinic we are using QAtrack+ for more than 6 months and we are gradually encoding all test there.
Recently I started some basic image analysis using the openCV when I run into your great work.
In my approach I would like to merge everything into QAtrack+ and install more and more packages within that product to have a single comprehensive solution.
I am not an experience programmer, but looking at various nice initiative I think some of them is just getting off track sooner or later, even though the initial idea/work is quite promising.
So making two /more python based solutions migrating or converging, to me this increases the visibility of each single project and discovering multiple point of view’s.
Probably a standalone GUI version would be more attractive for single users, but you can think that the GUI is already there for QAtrack+ (in a intranet/multiplatform way) thus with a single file upload function all the pylinac functionality could be assessed specifically at the local user’s frequency / tolerance tables.
So to my opinion this way both project would receive a boost in terms of future development.
(Even though Randle Taylor was less active for development since he left TOCC)
I agree that this is the up-and-coming thing in radiotherapy, so having a module for the phantom QA sounds like a good idea to me. I’ll have to research this a bit more to understand both the phantom and the procedure. Do you have or know of any images available to look at?