aboutsummaryrefslogtreecommitdiff
path: root/src/merge.c
AgeCommit message (Collapse)Author
2021-03-05Mass update of copyright dates in source code commentsThomas White
2020-03-05partialator: Gracefully handle negative and NaN sigma valuesThomas White
Of course, these should never happen.
2020-01-10Mass update of copyright datesThomas White
2019-07-17Modify residual some moreThomas White
2019-07-17Modify residual to give clearer minimum, particularly as function of profile ↵Thomas White
radius
2019-07-17Weight reflections according to how much they need to be scaled upThomas White
2019-07-17Allow correct_reflection() to correct anything, not just intensitiesThomas White
2019-06-21FormattingThomas White
About the comment: yes, it is. temp2 (T) has already been de-logged by this point.
2019-06-21Actually prevent the use of free-flagged reflections in refinementThomas White
Since we are now using residual() in a numerical minimisation, it's important that it uses either only free-flagged reflections, or only non-free-flagged ones.
2018-11-30Factorise correction of intensity for G, B, p and LThomas White
There were no fewer than 8 different places in the code where these factors needed to be applied. They all need to agree on conventions such as whether the intensities in the pattern should be multiplied or divided by G to "correct" them. They were not. This commit reduces the number of places to three: one function (actually two functions, so that the value before partiality can also be calculated consistently), plus log_residual and scale_one_crystal, where slightly different logarithmic versions are used instead.
2018-11-22Don't weight reflections by partiality in residualThomas White
Because we want under-prediction to be penalised just as much as over-prediction
2018-11-22log_residual: Actually set n_usedThomas White
2018-11-22Restore reflection weightingThomas White
2018-11-16partialator: Handle partial reflections properly in deltaCChalfThomas White
2018-11-14Merge branch 'tom/logmerge'Thomas White
2018-09-10Make sure reflection contribution list gets freedThomas White
2018-09-02FussinessThomas White
2018-08-30Save contribution information during mergeThomas White
2018-06-28Merge ln of intensity rather than intensities themselvesKen Beyerlein
2018-05-11Fix up scaling definitionsThomas White
2018-05-07Move residual() and log_residual() to merge.cThomas White
2018-05-02Revert 1/corr weighting during mergeThomas White
This restores the scaling behaviour (and performance) of version 0.6.3
2018-03-02Remove unnecessary pmodelsThomas White
2018-02-27Reduce the weight on reflections if they have to be scaled upThomas White
2018-02-27New rejection criteriaThomas White
2018-02-27Scaling fixesThomas White
2018-02-27Store scale factor in linear formThomas White
2018-02-27Remove message about number of reflections in mergeThomas White
2017-07-21Tidy up some warningsThomas White
2017-03-03partialator: Better handling of very small number of crystalsThomas White
2015-11-19Count reflections in merge and scalingThomas White
2015-11-18No free-flagged reflections in merge for scalingThomas White
2015-11-18Factorise get_locked_reflection()Thomas White
2015-11-18merge_intensities(): add use_weak optionThomas White
2015-08-31Trap attempt to merge zero crystalsThomas White
2015-07-16partialator: Don't merge reflections where isnan(corr)Thomas White
2015-05-19Debugging stuffThomas White
2015-05-19s/lsq_intensities/merge_intensities/Thomas White
2015-05-19Strip out old scaling codeThomas White
generated by cgit v1.2.3 (git 2.25.1) at 2024-11-16 15:00:09 +0000