Changes

* Fix for [https://{{gna.org/bugs/?bug link|24218 |text=bug #24218, the incorrect labelling of alignment tensors by the align_tensor.matrix_angles user function when a subset of tensors is specified]}}. The logic for the labels was expanded from being only for all tensors to handling subsets.* Bug fix for the [http://www.nmr-relax.com/manual/structure_read_pdb.html structure.read_pdb user function] [https://{{gna.org/bugs/?bug link|24300 |text=bug #24300, when the merge flag is True, and both multiple structures and multiple models are present, the structure.read_pdb user function would fail with a RelaxError]}}. The problem was that the molecule index was simply not being updated correctly.* Fix for [https://{{gna.org/bugs/?bug link|24601 |text=bug #24601, the failure of the optimisation of the R2eff dispersion model when peaks are missing from one spectrum]}}, as reported by Petr Padrta. To handle the missing data, the peak intensity keys are now checked for in the spin container peak_intensities data structure. This is both for the [[R2eff model|R2eff]] model optimisation as well as the data back-calculation. A warning is given when the key is missing. The [http://www.nmr-relax.com/api/4.0/specific_analyses.relax_disp.api.Relax_disp-class.html#base_data_loop relaxation dispersion base_data_loop() method] has been modified to now yield the spin ID string, as this is used in the warnings. In addition, the Grace plotting code in the relax library was also modified. When peak intensity keys are missing, some of the Grace plots will have no data. The code will now generate a plot for that data set, but detect the missing data and allow an empty plot to be created.