Visualization: Fix legend argument checking for waterfall/parameter/history plots

[m-philipps]

• before the fix a legend_error would be thrown when the user supplied the legends as a tuple
• Now handles all data types
• raises TypeError for single and multiple results
• raises ValueError in case of multiple results when lengths of results and legends do not match

[codecov-commenter]

Codecov Report

Merging #1139 (14ed5d2) into develop (883138a) will increase coverage by 1.30%.
The diff coverage is 46.15%.

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

@@             Coverage Diff             @@
##           develop    #1139      +/-   ##
===========================================
+ Coverage    83.06%   84.37%   +1.30%     
===========================================
  Files          148      148              
  Lines        11642    11649       +7     
===========================================
+ Hits          9671     9829     +158     
+ Misses        1971     1820     -151     

Files	Coverage Δ
pypesto/visualize/misc.py	73.55% <46.15%> (-1.89%)	⬇️

... and 8 files with indirect coverage changes

[PaulJonasJost]

Thanks for noticing, looks good aside from some slight changes .

[m-philipps]

@PaulJonasJost what do you think of the suggestion for raising different errors for legend type and length, also in the case of a single result?

[m-philipps]

ready to go, @plakrisenko

[plakrisenko]

if this is true it already means that 'List of results passed and list of labels do not have the same length.', right? As it's a multiple result case + only a string was passed as a legend

[m-philipps]

The string could have matching length, e.g. "ABC" for 3 results.

The idea was that a string of length != len(results) would throw an error, but a string with the appropriate number of characters would be used for the legend. We could also raise an error for any string, but this would not be consistent with the type hint for the legend argument of visualize.waterfall: legends: Optional[Union[Sequence[str], str]] = None, because a string is a Sequence.