Open SilverViking opened 4 years ago
I would go for the first option: let the user specify where the data is located, download and compress it. The source location will typically be over a network connection, which would result in significant delays in the backend.
Perhaps later, it would be more efficient to have a "Fast segment" button for very large datasets? This button would then download the data locally, convert it to a fixed format, segment it, discard the input data locally and write the results to a preferred location.
How are we going to handle the user's raw and annotated data: where is it going to be stored? Will we copy it (manually? via the biosegment frontend?) to a fixed location where biosegment can access it? Or should biosegment be able to access it at from its source location?