joining forces

[alexlib]

Hi @tmatsuzawa - I think we develop very similar things, though by different approaches. Your project though is by far more advanced. Mine is very primitive at the moment, see https://github.com/alexlib/pivpy

I suggest to join forces and get a package that works on Davis, but also other software packages, e.g. OpenPIV, etc. plus to get more functionality as in PIVMAT.

Regars
Alex

[tmatsuzawa]

Hi @alexlib. Happy to connect!
I like your use of xarray in your project. I had this idea as well to save metadata but ended up sticking to hdf5.

I'd be happy to work together to package that works on DaVis and OpenPIV, and perhaps PIVLab. Clearly, my package needs some cleaning before the public release, and I'd be more than happy to chat about the features such as making a movie from their outputs. My package already contains quite many features that I use in my research.

[tmatsuzawa]

A sample notebook I made 3 years ago (the actual notebook can be found in the same directory)
... it features basic functionalities. plotting, computing velocity gradient tensor, energy, enstrophy, autocorrelation, etc.
... Pretty much all algorithms have improved since then.

Documentation

[alexlib]

A sample notebook I made 3 years ago (the actual notebook can be found in the same directory) ... it features basic functionalities. plotting, computing velocity gradient tensor, energy, enstrophy, autocorrelation, etc. ... Pretty much all algorithms have improved since then.

Documentation

this looks great. I think the next steps would be:

• take the same data as presented in this notebook
• make it xarray, either using the pivpy reading or by adding a function that reads it in tflow and converts it to xarray that pivpy can read
• reimplement the existing functions from tflow to xarrays, thus building the option to work either in tflow or in pivpy seemingly
• I do not mind to have a single package, whatever we call it, as long as we save both sides time and get a better package. obviously your work is of higher Python and graphics quality, so we shall adopt your style 👍