Loop through symops:
apply each symop to the original asu,
then apply unit translations to minimize its distance to the original asu,
test if new asu intersects with the target 27-bricks-large box,
use it, if it does.
Keep in mind that asu is not always a brick, it is always a convex polyhedron in cctbx.
I do not see any good reason for doing all that.
I do not read cctbxbb.
On Jul 24, 2014, at 2:43 PM, Alastair Fyfe
Hi Marat, my apologies if this is a duplicate. From Pavel's comment it occurred to me you might not follow posts to the cctxbb Alastair ------
Many thanks for your replies.
Pavel : by "slow" I meant to (a) apply sym ops to asu and refer results back to unit cell thus obtaining a "full" unit cell (I expect "expand_to_p1()" does this?) (b) translate the cell and (c) find overlap with the 26-brick neighborhood of the starting brick. This should work but seems inefficient.
Marat: applying SymOps to the basis vectors seems a great idea. However, it wasn't clear to me how to combine sg SymOps to yield the 26 adjacent bricks. Does the map expansion code do this?