This issue is to collect ideas how to improve the library in its general design:
Ideas:
[ ] Data cube view is a parameter only to image_collection_cube, cube instances (except image_collection_cube) should not store its st reference but provide a function that derives the reference from input cube and maybe other parameters.
[ ] Make data cube objects copyable by copying all parameters (including input cubes)
[x] Create collection format JSON description presets
[ ] For some operations that must read adjacent chunks, a chunk cache with key(hash(graph), chunk_id) could improve the overall performance by reducing reads from disk
This issue is to collect ideas how to improve the library in its general design:
Ideas: