-
Notifications
You must be signed in to change notification settings - Fork 47
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
partiality_threshold=0.4 in cosym, symmetry and ttr #1470
Conversation
Previous default of 0.99 can be overzealous for particularly narrow wedges. Set to same default as dials.scale, dials.export and dials.merge.
Expose partiality_threshold parameter on command line. Previous default of 0.99 can be overzealous for particularly narrow wedges.
to reproduce the previous result (and test that the program supports setting it on the command line).
Codecov Report
@@ Coverage Diff @@
## master #1470 +/- ##
=======================================
Coverage 65.81% 65.82%
=======================================
Files 614 614
Lines 69410 69410
Branches 9508 9508
=======================================
+ Hits 45685 45691 +6
+ Misses 21887 21884 -3
+ Partials 1838 1835 -3 |
For |
Perhaps I'm missing something, but I don't understand why this would be the case? The default https://github.com/dials/dials/blob/master/util/filter_reflections.py#L457 |
With
With this PR:
I.e. with this PR, fewer reflections are removed (none in this case) for being below the partiality threshold, and more reflections in total are used in refinement. |
I see, thanks. I didn't realise anything was filtered previously in |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It makes sense that these tools use the same threshold for post-scaling reflection selection. I did wonder whether it would make sense for filter_reflection_table
to set a flag, but I think that can be a separate discussion.
* Default partiality_threshold=0.4 in cosym and symmetry Previous default of 0.99 can be overzealous for particularly narrow wedges. Set to same default as dials.scale, dials.export and dials.merge. * Default partiality_threshold=0.4 in ttr Expose partiality_threshold parameter on command line. Previous default of 0.99 can be overzealous for particularly narrow wedges.
3.2 Branch releases will now use a fixed conda environment. This release is the first to use the same versions of all dependencies as 3.2.0. Bugfixes -------- - ``dials.symmetry``, ``dials.cosym`` and ``dials.two_theta_refine``: Lowered default partiality_threshold from ``0.99`` to to ``0.4``. The previous default could occasionally result in too many reflections being rejected for particularly narrow wedges. (#1470) - ``dials.stills_process`` Improve performance when using MPI by avoiding unnecessary log file writing (#1471) - ``dials.scale``: Fix scaling statistics output of r_anom data. (#1478)
3.2 Branch releases will now use a fixed conda environment. This release is the first to use the same versions of all dependencies as 3.2.0. Bugfixes -------- - ``dials.symmetry``, ``dials.cosym`` and ``dials.two_theta_refine``: Lowered default partiality_threshold from ``0.99`` to to ``0.4``. The previous default could occasionally result in too many reflections being rejected for particularly narrow wedges. (#1470) - ``dials.stills_process`` Improve performance when using MPI by avoiding unnecessary log file writing (#1471) - ``dials.scale``: Fix scaling statistics output of r_anom data. (#1478)
Set the default partiality_threshold to 0.4 in dials.cosym, dials.symmetry and dials.two_theta_refine. The previous default of 0.99 could occasionally be overzealous in rejecting reflections, especially with particularly narrow wedges. This issue was identified by errors encountered running xia2.multiplex on James Holton's "Micro-focus Data Processing Challenge" dataset.