Collecting at T2 with T1 positions

When fitting data at two different temperatures, it is important that the data at T2 have the same relative position respect to the spot center, in reciprocal space, as the data at T1.

In the tutorial on the MgO case we have explained how to use the cloning script. This script take as input a file collected at T1 which has been generated to contain, beside the data, also the the information on how to clone by interpolation the same Q points for the data at T2.

This generation was triggered by specifyng the DELTA dictionary. This dictionary contains the difference in alignement between T2 and T1. Here we report the concerned part of the MgO tutorial, with additional information to do the cloning also with on-Bragg recentering :

Cloning by a mere geometrical alignement

Suppose that you have aligned both datasets at T1 and T2 with the parameter files tds2el_90.par aand ../120/tds2el_120.par You can obtain the differences in alignement with this command

tds2el_getDelta_v1 tds2el_90.par  ../120/tds2el_120.par

to get this output

DELTA={
 "r1":-9.99999999973e-07,
}

in the case they only differe by a minimal amount in r1 variable. All the other variables which have changed between T1 and T2 will appear in DELTA.

the DELTA dictionary contains the differences which, added to the 90 K parameters, give the 120K ones. We create now a new input file by adding the dictionary definition at the end of tds2el_90.par (copy and past, at the end of the par file, DELTA={...} as from the output of tds2el_getDelta) and redo the harvesting on collection.h5. ( it is already done if you have used the par file as it is) Now, thank to the DELTA option, the collected file will contain, beside the usual harvested data, additional informations that will be used to interpolate the 120K dataset to obtain interpolated datapoints that correspond to the same Q points at which we have the data points in the 90 K case.

We have now, in collection.h5 file, the interpolation informations for each data point: they are obtained calculating for each 90K data point the combination of pixels 120 K data that have to be taken in order to reproduce the same 90 K Q point using the 120 K dataset whose alignement differs by DELTA.

With this information we can clone collection.h5 using data for the 120 K dataset. We use for this the *clone_120.inp* input file

images_prefix         =  "../../data/120/mgo__0001p_0"
filter_file           ="../filter.edf"
file2becloned = "../90/collection.h5"
clonefile = "collection_120_cloned.h5"
normalisation_file    = "../90/normalisation_120.txt"
threshold = 1000
badradius = 0

The command is, to be given in analysis/120

tds2el_cloneinterp_v1 clone_120.inp

here images_prefix points to 120 K directory, the interpolating information is read from file2becloned, the result goes to clonefile. The interpolation is applied on the 120 K datapoints, after normalisation by normalisation_file. The regions above threshold points are not used in the fit. If badradius is zero a whole image is discarded if one of its pixels is higher that threshold. Otherwise for each pixel above-thrshold all pixels within badradius are rejected.

Cloning by considering on-Bragg centering

We have also seen in the tutorial on the Calcite case that when the Bragg spot positions are precise enough their position (calculated as the baricenter of those points inside Q_min with intensity higher than threshold) can be used to redefine the center of each spot.

In this case beside providing DELTA a list of two centering file must be provided. The workflow is the following

  • a harvesting for T1

  • a harvesting for T2

  • then a recentered havesting for T1 havong set in the input

    DELTA={ "r1":0.000000001, or something else }
centering_file =["collection_for_T1.h5", "collection_for_T2.h5"]

the generated file can then be cloned by giving its name as file2becloned to the cloning script.