Closed cyril-data closed 1 year ago
Thank you both! I have one quick question about how this works in practice. With this new updated, we have 1 block == 1 activity (instead of 1 command == 1 activity), can you tell me how are inputs/outputs dealt with? Specifically: 1- are we still defining inputs/outputs at the command level (and then the block knows which commands are in that block and can deduce its inputs/ouptus) 2- is there a new mecanism at the block level 3- another solution?
Thanks!
Hi @cmaumet ,
This is exactly the hypothesis 1- are we still defining inputs/outputs at the command level (and then the block knows which commands are in that block and can deduce its inputs/ouptuts)
The activities and their associated entities (inputs or outputs), are read and assembled in a graph as usual. But in parallel, at the reading procedure of the afni log file, we have associated each activity to a block : we know which commands are in each block. And so, we can deduce the block inputs and ouptuts and build a new block graph, deleting all intermediate entities inside the block.
Cheers
Awesome thanks @cyril-data!
This PR concatenates activities in blocs define by the start of "# ====== ..." in logs of afni.