Added some retry/exponential backoff code. Started an NDVI run of US and it will now start on the last successful tile. Set concurrency in open_dataset, which fixes my concurrency error but it is very slow now. Trying to write exponential backoff code that can catch this error and contnue to retry (or poke) Google. Problem is the my exponential backoff function is wrapped around to_dataframe() which in theory is supposed to be where the Earth Engine call happens but it's not working.

Author: adrianom-gh

.gitignore

@@ -52,3 +52,14 @@ geometry/US.shp
 geometry/US.prj
 geometry/US.dbf
 geometry/US.cpg
+demos/US_NDVI_Tiles_30m_40MIL
+
+src/robustraster/__pycache__/ee_grid_tiles.cpython-310.pyc
+src/robustraster/__pycache__/format_time.cpython-310.pyc
+src/robustraster/__pycache__/hooks.cpython-310.pyc
+src/robustraster/__pycache__/raster_export_manager_batched.cpython-310.pyc
+src/robustraster/__pycache__/reports_setup.cpython-310.pyc
+src/robustraster/__pycache__/run_manager.cpython-310.pyc
+src/robustraster/__pycache__/vector_export_manager.cpython-310.pyc
+src/robustraster/__pycache__/vrt_export_manager.cpython-310.pyc
+src/robustraster/__pycache__/dask_docker_cluster_manager.cpython-310.pyc