Publication of studies with errors should be blocked in curation interface

[ljwh2]

Recently we had some examples where studies were published with no reported trait or discovery sample information (mentioned here: [https://github.com//issues/144]). Publication of studies with these major errors should be prohibited in the curation interface, in the same way as publishing with no EFO trait is prohibited.

The full list of errors which should block publication (for the ones in italics the functionality already exists):
No EFO trait
No genotyping technology
No Reported trait
No text in Discovery sample information
No structured ancestry

The expected behaviour is:

1. Curator sets status to "publish study" (either in "Curator information" tab or from Curation homepage)
2. Curator is presented with error message: "Study: Smith J, pubmed = 12345678 requires review before publication because"
   followed by
   "no reported trait assigned"
   "no Discovery sample information entered"
   "no structured ancestry assigned"
   This error message lists all errors blocking publication.

See attached screenshot for current error message, generated by attempting to publish a study with no data.

[ljwh2]

Note, curators think this feature used to exist for reported trait

[ljwh2]

There is one additional place where the status can be set to published, which is the publication view:

https://www.ebi.ac.uk/gwas/curation/publication/32693751

Attempting to publishing from here does not currently generate any error messages (including EFO and genotyping technology shown in screenshot above). This status change should generate the same error messages as described above.

[JalMacArthur]

Discussed on 19th Aug - decided this needs further investigation to determine what QC of data is required and at what points. Taking into account submitted and curated data. Suggest meeting and discussion between Laura, Jackie, Tudor, Yomi.