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 <afyfe(a)soe.ucsc.edu> wrote:
>> 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?
>>
>>
>
