harmonize 'nan' vs empty strings vs ?

[WardDeb]

mismatching parkour dfs vs samplesheets give rise to empty strings (rather then nans) in the df. This can result in the issue below (though not on re-runs ?).

Todo:

• harmonize 'empty' in dfs to np.NAN across entire workflow
• have a check / issue warnings when samplesheet & parkour are not matching (especially in forward direction).
• deal correctly with empties upon multiqc & email creation.

File "/home/pipegrp/src/github.com/maxplanck-ie/dissectBCL/src/dissectBCL/fakeNews.py", line 217, in multiQC_yaml                                                           
    reqDepth = float(                                                                  
ValueError: could not convert string to float: ''                                                                                                                             

Originally posted by @dkoppstein in #92 (comment)