BandsData: problems with k-point labels.

[yakutovicha]

It is not enough to correctly import a set of k-points and bands into a BandsData object to be able to use it properly. If k-points labels are not specified - the bands object will complain:

/home/epfl/work/venv-lsmo/local/lib/python2.7/site-packages/aiida/orm/data/array/bands.py in _matplotlib_get_dict(self, main_file_name, comments, title, legend, legend2, y_max_lim, y_min_lim, y_origin, prettify_format, **kwargs)
    805         labels = plot_info['labels']
    806         # prepare xticks labels
--> 807         tick_pos, tick_labels = zip(*labels)
    808 
    809         #all_data['bands'] = the_bands.tolist()

ValueError: need more than 0 values to unpack

if I add (at least one) labels: it works

In [9]: n.labels =[(0,'G')]

In [10]: n.show_mpl()

In [11]: 

another problem occurs if one does not include the labels of the very first and very last k-point:

In [11]: n2.labels =[(10,'A'), (20, 'B')] # total number is 39

In [12]: n2.show_mpl()

In [13]: 

as one can see in the figure below, the k-points 0-9 and 21-38 were cut:

And the last thing is the lack of documentation. Without help I would spend a lot more time trying to figure out the label-related logic.

Therefore the suggestions to improve would be:

• - fix empty labels array bug
• - fix problem with cutting k-points
• - add documentation on how to create and visualize BandsData object

[sphuber]

Fixed in PR #2492