Always use linking ISSN when it exists

[paulalbert1]

Background

Looking up journal-specific metrics is easiest when we use linking ISSN to fill the ISSN column in the personArticle table.

Problem

In actuality, this code for import from s3 (and its counterpart for DynamoDB) prefers e-ISSN even when there's a linking ISSN.

ReCiter-MachineLearning-Analysis/S3_AnalysisOutput_upload.py

Line 257 in 39306cb

if 'issn' in items[i]['reCiterArticleFeatures'][j]:

For example, here's the output for szd3005 and 33476307.

{
    "personIdentifier": "szd3005",
    "dateAdded": "2021-01-26T20:33:07.606+00:00",
    "dateUpdated": "2021-01-26T20:33:07.606+00:00",
    "reCiterArticleFeatures": [
        {
            "pmid": 33476307,
            "journalTitleVerbose": "The Journal of clinical investigation",
            "issn": [
                {
                    "issntype": "Electronic",
                    "issn": "1558-8238"
                },
                {
                    "issntype": "Linking",
                    "issn": "0021-9738"
                }
            ],

To populate ReCiterDB, we're pulling electronic ISSN as opposed to linking ISSN. Unfortunately, this is quite common.

Fix

The relevant Python scripts should preferentially use linking ISSN whenever it exists.