no indication of problem lines in badly-formed GFFs

[joelb123]

Running gffpandas over a badly-formed GFF doen't give any indication where the problem occurred. Line numbers in ValueError messages would be invaluable. To accomplish this, you'll have to refactor a bit. Consider adding a
line-number column to the data frame for temporary internal-only use and doing onversions in df.iterrows() or df.groupby() using try/except rather than in apply(). apply() fails silently by design, which is good for some situations but not this one.

[joelb123]

I'll also note that even a totally spurious text file (say, a FASTA instead of a GFF) does not produce an error, just no features. GFFpandas is often embedded a few layers deep, and raising an exception with a message indicating the file giving the problem would be much appreciated.