Reads in TurboSM data files (.tsm, .tbn), bins to an image size that the original C++-only PhotoZ can handle, translates or fills in metadata, and writes a legacy ZDA file
Clone and install conda environment.
git clone https://github.com/john-judge/TSM-to-ZDA.git
cd TSM-to-ZDA
conda env create -f environment.yml -n TSM_to_ZDA
conda activate TSM_to_ZDA
jupyter notebookJupyter will open in browser. Open the ipynb file.
Adjust filename and directory as needed. Default binning and cropping should work well for DaVinci recordings and original PhotoZ.
As of 5/27/22, output is simply hardcoded as output.zda in the current directory.
https://github.com/john-judge/PhotoZ_upgrades/tree/load-hacked-zda is a version of PhotoZ created for testing with this script and is known to be compatible. This version ignores some PhotoZ validation such as version and bit-depth, which do not need to be written correctly for the data to load properly. This version also includes an initially high but adjustable binning setting in PhotoZ to avoid performance issues.
PhotoZ only plays nicely with ZDA files containing square images (i.e. where width equals height).
Binning is high for large images to avoid performance and memory issues in PhotoZ.