Update ingestMetadataByDOI process

[rickjohnson]

• use parameterizable list of CSV files
• Update to use NormedPerson class and related methods
• Update to use NormedPublication class and related methods

[jeremyf]

I'm not understanding this issue. In the current ingestMetadataByDoi.ts I'm not seeing any hard-coded CSV files.

Similarly, I'm uncertain what you mean by "update to use NormedPerson" and "update to use NormedPublication". There's a "SimplifiedPerson" interface that's intertwined in the code.

What I'm seeing in the ingest process is a lot of JS mapping CSV to JSON with some munging; And then more mapping.

There's a lot of programmatic mapping going on, and it's really hard to track those concepts in the ingest (because we're working with lots of loose Hashes being mapped repeatedly). And try as I might, I can't seem to track all of these programmatic transformations (in part because it's a tremendous amount of data cross walking done in code).

What I'm beginning to wonder is if the implementation of much of what we have would be better served with CSV dumps into SQL tables; Each CSV file with common headers (no need to overly normalize) would be its own table. Then, we could run SQL commands to transform and copy those tables into a normalized table.

In other words, I'm really struggling with feeling helpful on this project. In part because I'm also wrestling with anxiety at the timelines, unshared obligations/expectations, and disjoint communication channels around this project.