Sorry Nick,

please let me know - I can pull it back or tweak as needed...

Cheers,

Oleg.

---

From: cctbxbb-bounces@phenix-online.org <cctbxbb-bounces@phenix-online.org> on behalf of Nicholas Sauter <[email protected]>
Sent: 05 September 2018 16:37:53
To: cctbx mailing list
Cc: [email protected]
Subject: Re: [cctbxbb] [git/cctbx] master: re-defining the shift/est convergence limit in terms of crystallographic rather than independent parameters to make it easier to relate to the convergence reported in the CIF; also caching the shifts for further analys...

 

I'm worried about making changes like this to a fundamental toolbox component like normal equations solving.  We use this for processing xfel data, and would like the opportunity to test any changes through a pull request, prior to having them committed to the main branch.

Nick Sauter

Nicholas K. Sauter, Ph. D.
Senior Scientist, Molecular Biophysics & Integrated Bioimaging Division

Lawrence Berkeley National Laboratory
1 Cyclotron Rd., Bldg. 33R0345
Berkeley, CA 94720
(510) 486-5713

On Wed, Sep 5, 2018 at 5:29 AM, CCTBX commit <diamondlightsource.jenkins@gmail.com> wrote:

Repository : ssh://g18-sc-serv-04.diamond.ac.uk/cctbx
On branch  : master

---

commit 9339fd5f7d240f0d6f85dd57ce42e796d19608f1
Author: pcxod <[email protected]>
Date:   Wed Sep 5 13:29:29 2018 +0100

    re-defining the shift/est convergence limit in terms of crystallographic rather than independent parameters to make it easier to relate to the convergence reported in the CIF; also caching the shifts for further analysis and reporting;

---

9339fd5f7d240f0d6f85dd57ce42e796d19608f1
scitbx/lstbx/normal_eqns_solving.py | 17 +++++++++++------
1 file changed, 11 insertions(+), 6 deletions(-)

diff --git a/scitbx/lstbx/normal_eqns_solving.py b/scitbx/lstbx/normal_eqns_solving.py
index 660304eab..45f070622 100644
--- a/scitbx/lstbx/normal_eqns_solving.py
+++ b/scitbx/lstbx/normal_eqns_solving.py
@@ -140,15 +140,20 @@ class iterations(object):
     a = self.non_linear_ls.normal_matrix_packed_u()
     a.matrix_packed_u_diagonal_add_in_place(value*a.matrix_packed_u_diagonal())

-  def do_scale_shifts(self, max_shift_over_esd):
+  def do_scale_shifts(self, limit_shift_over_su):
     x = self.non_linear_ls.step()
     esd = self.non_linear_ls.covariance_matrix().matrix_packed_u_diagonal()
-    x_over_esd = flex.abs(x/flex.sqrt(esd))
-    max_val = flex.max(x_over_esd)
-    if max_val < self.convergence_as_shift_over_esd:
+    ls_shifts_over_su = flex.abs(x/flex.sqrt(esd))
+    #max shift for the LS
+    self.max_ls_shift_over_su = flex.max(ls_shifts_over_su)
+    jac_tr = self.non_linear_ls.actual.\
+      reparametrisation.jacobian_transpose_matching_grad_fc()
+    self.shifts_over_su = jac_tr.transpose() * ls_shifts_over_su
+    self.max_shift_over_su = flex.max(self.shifts_over_su)
+    if self.max_shift_over_su < self.convergence_as_shift_over_esd:
       return True
-    if max_val > max_shift_over_esd:
-      shift_scale = max_shift_over_esd/max_val
+    if self.max_ls_shift_over_su > limit_shift_over_su:
+      shift_scale = limit_shift_over_su/self.max_ls_shift_over_su
       x *= shift_scale
     return False

---

To unsubscribe from the CCTBX-COMMIT list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCTBX-COMMIT&A=1

_______________________________________________
cctbxbb mailing list
[email protected]
http://phenix-online.org/mailman/listinfo/cctbxbb