Empty rows should be dropped

[joemull]

The CSV update routine tries to process rows that do not have any data, like the last line of this import:

Janeway ID,Article title,Article abstract,Keywords,Rights,Licence,Language,Peer reviewed (Y/N),Issue title,Author salutation,Author given name,Author middle name,Author surname,Author suffix,Author email,Author ORCID,Author institution,Author department,Author biography,Author is primary (Y/N),Author is corporate (Y/N),DOI,DOI (URL form),Date accepted,Date published,Article number,First page,Last page,Page numbers (custom),Competing interests,Article section,Stage,File import identifier,Journal code,Journal title override,ISSN override,Volume number,Issue number,Issue pub date,PDF URI
,"The Palimpsest, vol.8 no.2, February 1927",,,Material in the public domain. No restrictions on use.,publicDomain,eng,N,,,,,,,,,,,,,,,,,1927-02-01,,,,57-88,,Full Issue,Published,,palimpsest,,,8,2,1927-02-01,https://sdrc.lib.uiowa.edu/Uploads/temp/Pal1927_February_V8_N2.pdf
,A Capital in Name Only,,,Material in the public domain. No restrictions on use.,publicDomain,eng,N,,,John,Ely,Briggs,,,,,,,N,N,,,,1927-02-01,,,,57-70,,Article,Published,,palimpsest,,,8,2,1927-02-01,https://sdrc.lib.uiowa.edu/Uploads/temp/Pal1927_February_V8_N2-02.pdf
,Flashlights on Vicksburg,,,Material in the public domain. No restrictions on use.,publicDomain,eng,N,,,Bessie,L,Lyon,,,,,,,N,N,,,,1927-02-01,,,,71-80,,Article,Published,,palimpsest,,,8,2,1927-02-01,https://sdrc.lib.uiowa.edu/Uploads/temp/Pal1927_February_V8_N2-03.pdf
,Comment by the Editor: A Lifetime Ago,,,Material in the public domain. No restrictions on use.,publicDomain,eng,N,,,John,Ely,Briggs,,,,,,,N,N,,,,1927-02-01,,,,81-88,,Article,Published,,palimpsest,,,8,2,1927-02-01,https://sdrc.lib.uiowa.edu/Uploads/temp/Pal1927_February_V8_N2-04.pdf
,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

This causes an error message "No journal found" on the affected row, but the import works for other rows, which is confusing to users.

We should check for empty data in prepare_reader_rows and not add empty rows to article_groups.