restore some test cases from mkresch branch

[yxqd]

The data are in pkl format. Convert them to hdf5 histogram files.

[yxqd]

Doug suggested using Al data as the next example. That is a good idea since it is almost pure coherent.

[yxqd]

Seems working for Al now.

Iteration of DOS:

Last iteration S(E):

Last iteration S(Q,E):

[yxqd]

Here are updated results for Vanadium data:

Iteration of DOS:

Last iteration S(E):

Last iteration S(Q,E):

[yxqd]

What happen if a large, incorrect C_ms was used? The following is the result for C_ms=1. for the V data. Large residuals are observed.

The DOS iteration: