relax SVN archive

Attempt at fixing bug #21080 (https://gna.org/bugs/?21080).

This was reported by Troels Linnet.

The problem is a standard GUI problem. The text from a wxPython GUI is a unicode string. But relax
requires standard strings. Therefore the gui.string_conv.gui_to_str() function needs to be used on
the return value of the GUI element GetValue() method, but this call was forgotten.

Fix for the TP02 dispersion model.

The rotating frame tilt angle for this model is calculated from the population averaged chemical
shift and not the equal weighted average.

Turned off clustering in the Relax_disp.test_r1rho_off_res_fixed_time_tp02 system test.

This speeds the test up by more than half.

Fixes for the parameter checks in the Relax_disp.test_r1rho_off_res_fixed_time_tp02 system test.

The parameter values had not been updated from when the test was copied from one of the other tests.

Fix for the synthetic data for the TP02 dispersion model.

The nitrogen chemical shift was not converted from ppm to rad/s before being used to calculate the
offsets.

Comment fix for the loop_cluster() function.

Fix to allow R1 data to be randomised for Monte Carlo simulations for off-resonance R1rho data.

This is a temporary kludge for the dispersion analysis and needs to be replaced by a cleaner
solution via the base_data_loop() method.

Fix for the Relax_disp.test_bug_20889_multi_col_peak_list GUI test.

The spectrum.read_intensities user function no longer has 'heteronuc' and 'proton' arguments.

Added checks to the specific_analyses.relax_disp.disp_data.return_r1_data() function.

This is to help indicate to the user when data is missing.

The dispersion specific code is now assembling the R1 data and passing it to the target function.

The new specific_analyses.relax_disp.disp_data.return_r1_data() function creates a data structure
holding all the R1 data. This is used in the off-resonance R1rho experiments.