Improve plot data handling

[janssenhenning]

Also see #26

This PR implements a way to provide data to the plotting functions in many different ways while keeping the plotting functions consistent and simple.

A class PlotData is added, which acts as a light wrapper around mappings of many kinds (dicts, dataframe, ColumnDataSources), which can be used to easily handle multiple datasets. Suppose we have a dict data with a key x and y1 and y2 and we want to plot the y keys both against x

from masci_tools.vis.data import PlotData

p = PlotData(data, x='x', y=['y1','y2'])

for entry, source in p.items():
    print('x', entry.x, source[entry.x]) 
    print('y', entry.y, source[entry.y])

The above example will yield two items with x vs y1 and x vs. y2, source is the dict data in both cases. Notice that the chosen names x and y are completely arbitrary. With the initialization we define how our data columns should be named.

PlotData also has some convenience methods for mutating/copying/getting information about the data

In addition a function process_data_arguments is implemented, which can either take a mapping and kwargs to directly Initialize the PlotData as above. But it is also possible to provide only the kwargs with the wanted data in arrays/lists/series, which are then split up and put into dicts to initialize a PlotData object

[codecov]

Codecov Report

Merging #54 (a5ed43d) into develop (e4d9ea2) will increase coverage by 3.03%.
The diff coverage is 88.87%.

❗ Current head a5ed43d differs from pull request most recent head d722bc9. Consider uploading reports for the commit d722bc9 to get more accurate results

@@             Coverage Diff             @@
##           develop      #54      +/-   ##
===========================================
+ Coverage    78.15%   81.18%   +3.03%     
===========================================
  Files          100      102       +2     
  Lines        18568    20414    +1846     
===========================================
+ Hits         14512    16574    +2062     
+ Misses        4056     3840     -216     

Impacted Files	Coverage Δ
masci_tools/vis/bokeh_plotter.py	70.30% <56.33%> (+38.40%)	⬆️
masci_tools/vis/bokeh_plots.py	49.54% <68.42%> (+41.66%)	⬆️
masci_tools/vis/matplotlib_plotter.py	85.77% <68.42%> (-1.62%)	⬇️
masci_tools/vis/fleur.py	81.38% <87.17%> (+7.64%)	⬆️
masci_tools/vis/data.py	87.73% <87.73%> (ø)
masci_tools/vis/plot_methods.py	62.81% <88.41%> (+2.07%)	⬆️
masci_tools/vis/parameters.py	91.56% <91.66%> (+0.68%)	⬆️
tests/conftest.py	95.61% <94.00%> (-0.34%)	⬇️
tests/test_plot_data.py	99.41% <99.41%> (ø)
masci_tools/io/common_functions.py	93.47% <100.00%> (-1.45%)	⬇️
... and 30 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update e4d9ea2...d722bc9. Read the comment docs.

[janssenhenning]

TODO:

• Investigate how PlotData behaves for nan values (probably some extra care has to be taken)
• Implement Unit tests for PlotData
• bokeh dependency is not optional at the moment (import in data.py)