Hi Matthew,

TL;DR: You should stop spending time on the effort of getting MGET working on your Mac. But there are some other possibilities.

Thanks for your interest in MGET. Sadly, at this time, no it is not possible to easily use MGET on a Mac. MGET started as an extension to ArcGIS Desktop, which was and still is a Windows-only app. At the time, the best way forward was to take some Windows dependencies. We have slowly been removing those with an eye to multiplatform integration with QGIS, now that QGIS supports plugins. But there is still some distance to go and it is a back-burner project.

MGET’s Cayula-Cornillon algorithm is implemented in C++ here. The main function starts on line 227. (Although it is a C++ file, it is mainly C code.) If you are skilled in C and R, you might be able to adapt this code to your own R package. But it will be a tedious experience, as the code is somewhat tightly integrated with Python’s types and numpy ABI. If you do not have a lot of C experience, I would not advise this route. But if you do try it, feel free to ask me questions. (No promises I can help if they are very time consuming. :-)

I noticed you are at URI. There are some leaders in front detection there—Igor Belkin, Dave Ullman, even Peter Cornillon himself. It might be worthwhile to reach out to them to see if they’re interested in collaborating or had some code that would be easier to adapt to your situation. If you like I may be able to dig up Cayula’s original RatFor (Rational Fortran) code that Dave Ullman sent to me many years ago. But it is quite messy (no disrespect to Cayula) and would be even more of a challenge to work with.

I may also have a MATLAB implementation of the algorithm by Alistair Hobday, but it would be best to contact him directly and see if he’s still working with it or has something better.

Finally, you could consider the Canny algorithm instead. There is an R implemention in the biOps package. I have not used it in R but have had good success with it applying a MATLAB implementation to front detection problems. The challenge will be dealing with clouds. Most generic edge detection code I have come across assumes that there are no “missing pixels” from the image and ends up drawing edges around all of these regions. This is obviously not what you want when working with oceanographic data.

If MATLAB is useful to you, I wrote a version of Canny that handles the missing pixels problem.

Hope this helps,

Jason

I want to use MGET to use the Cayula-Cornillon SST front algorithm, but I am limited to a Mac with R and Python. Is there any way for me to use MGET? I've searched some of your documentation and it seems I would need DCOM which doesn't seem to be available for Mac? I'm not necessarily asking for step-by-step directions, I just need to know if it is possible or if I should stop spending time on the effort.