Slight update on creating Finngen studies.

[DSuveges]

There's a minor update in scripts/make_FINNGEN_study_table.py:

• Table is read from JSON directly from URL, no need to create a local copy and re-format via jq. (hence there's a little update in the snakefile.)
• If phenostring is not present, or it's an empty string, phenocode is used to describe the phenotype.
• Other updates are trivial.

Resulting json line for the problematic study descibed in #2585:

{
  "study_id": "FINNGEN_R6_I9_HEARTFAIL_AND_CHD",
  "pmid": "",
  "pub_date": "2022-01-24",
  "pub_journal": "",
  "pub_title": "",
  "pub_author": "FINNGEN_R6",
  "trait_reported": "I9_HEARTFAIL_AND_CHD",
  "trait_efos": null,
  "ancestry_initial": "European=206656",
  "ancestry_replication": "",
  "n_initial": 206656,
  "n_cases": 8876,
  "n_replication": 0
}

[ireneisdoomed]

Thank you so much for the fix @DSuveges

I've run the tables generation pipeline and I can confirm that everything runs smoothly and the bug reported in #2585 is fixed.
I9_HEARTFAIL_AND_CHD was the only study where the reported trait was an empty string, therefore having the problem.
As proposed, the reported trait has been successfully coalesced with the phenotype code so that the ETL will not dropped this record when filtering out those with trait_reported == null

>>> studies[studies['study_id'] == 'FINNGEN_R6_I9_HEARTFAIL_AND_CHD'].iloc[0]
study_id                FINNGEN_R6_I9_HEARTFAIL_AND_CHD
pmid
pub_date                                     2022-01-24
pub_journal
pub_title
pub_author                                   FINNGEN_R6
trait_reported                     I9_HEARTFAIL_AND_CHD
trait_efos                                         None
ancestry_initial                      [European=245732]
ancestry_replication                                 []
n_initial                                        245732
n_replication                                       0.0
n_cases                                         11034.0
num_assoc_loci                                        0
has_sumstats                                       True