Currently, users can/have to specify templates for each stage of processing. In practice, it is unlikely that anyone would intend to use multiple at various stages of process for the vast majority of experiments — almost certainly not without it being a key avenue of exploration. With the present configuration, there is too much room for error that we are noticing even when building configs internally.
Proposed feature
Global template specification. Something nearly as simple as "MNI2006", e.g., that would internally map that template and its derivatives to the resource pool where appropriate. Individual derivatives (e.g., template tissue maps) could also be overridden manually, taking their place in the prepopulated list. Finally, nodes could still override much in the same way. However, they will all inherit the decision from the data-config specified template, by default.
Related problem
Currently, users can/have to specify templates for each stage of processing. In practice, it is unlikely that anyone would intend to use multiple at various stages of process for the vast majority of experiments — almost certainly not without it being a key avenue of exploration. With the present configuration, there is too much room for error that we are noticing even when building configs internally.
Proposed feature
Global template specification. Something nearly as simple as "MNI2006", e.g., that would internally map that template and its derivatives to the resource pool where appropriate. Individual derivatives (e.g., template tissue maps) could also be overridden manually, taking their place in the prepopulated list. Finally, nodes could still override much in the same way. However, they will all inherit the decision from the data-config specified template, by default.
Acceptance criteria
The above.
Alternatives
No response
Additional context
No response