Currently we make use of a hard-coded file (dti_pipeline_regions.txt) at import-time.
This file is specific for the connectivity_96; so any other import from CSV fails (with message "Invalid Connectivity Row size! 74 != 96 at row 1"
One possible solution:
- make the "regions_order" part of the datatype - optional
- only the connectivity 96 will have, for now, this field populated
- at import from CSV, in case the regions_order att is missing, consider a default order 1....N
Let the canonical region order be the one in the reference connectivity (as specified in the centers.txt).
- if the CSV files do not contain a header row, then assume that the order is the canonical one
- if the CSV files contain a header row, that row must contain ordinal numbers.
We will permute the columns/rows so that the ordinal numbers will be in ascending order.
Then we assume that the sorted file is in the canonical order.
No extra mapping file is required. To adapt the ordering of the csv matrices to the reference connectivity add a header row with ordinal numbers.