My suggestion for tackling these would to create a pickled GEO or DSM object to use for testing, start with modifying some of the simpler functions and work backwards toward the read functions. Maybe?
- Read functions - will need to create objects
- Shape creation functions
- Conversion functions - I think all will just be changes to accessing attributes and debug images
- Analysis functions - I think these all currently pass the
spectral_object.array_data to rasterstats, so maybe will be simple
- Helpers
- Others
My suggestion for tackling these would to create a pickled GEO or DSM object to use for testing, start with modifying some of the simpler functions and work backwards toward the read functions. Maybe?
read.geotifread.netcdfauto_grid- just needed for attributes and debuggrid_from_coords- just needed for attributes and debugInteractiveShapesclass - will need changes throughoutpointsshapesto_roispectral_object.array_datato rasterstats, so maybe will be simplecolor- uses zonal stats after separating channels and doing colorspace conversioncoverage- only passes the binary mask to rasterstats, image is only used for metadata and debugspectral- this one might be more complicated, uses a real spectral object created by plantcv. Might just need a wrapper for thepcv.spectral_indexfunctions that takes our image class and returns an actual spectral object. Or returns another geo image?height_percentile- both height functions use the array_data directlyheight_subtraction_transform_geojson_crs_show_geojson_plot_bounds_pseudocoloredtransform_polygonscenter_grid_rois