Configure Bulkrax

[jillpe]

No description provided.

[bkiahstroud]

Attempting to import the sample CSV that the client provided caused the importer to fail with this error:

CSV::InvalidEncodingError - Invalid byte sequence in UTF-8 in line 3.

After some digging, I'm pretty sure the file is encoded in WINDOWS-1252 (primarily due to seeing <92> in vim instead of '). Running the following seems to have resolved the importer error:

iconv -f WINDOWS-1252 -t UTF-8 input.csv > output.csv

---

[UPDATE] -- We'll need to override this Bulkrax::CsvEntry option from utf-8 to WINDOWS-1252:UTF-8. This will read the file as WINDOWS-1252 encoded and then transcode it to UTF-8 (i.e. the default internal encoding for Ruby)

[UPDATE 2] -- The client was able to encode their CSVs as UTF-8, so this is no longer necessary

[bkiahstroud]

While configuring Bulkrax's field mappings using the CSVs provided by the client, it became apparent that WCU's field mappings conflict with other existing HykuUP tenants. Because of this, @orangewolf and I decided that the best short-term¹ solution was to add a new Hyku Account setting that would allow for per-tenant field mappings. That Hyku PR can be found here:

• Add per-tenant Bulkrax field mappings samvera/hyku#2384

Footnotes

1. A better long-term solution would be if Bulkrax allowed you to edit a given importer's field mappings. That feature would live entirely within Bulkrax and thus wouldn't have to "bleed" into Hyku. However, that's too much of a lift for the scope of this ticket ↩

[bkiahstroud]

The Hyku feature that allows for per-tenant field mappings has been merged into Hyku main. The next steps for this ticket are to:

1. Update the submodule in HykuUP knapsack to latest Hyku main
2. Deploy to prod
3. Configure the WCU and UNCA tenant field mappings

The Mobius field mappings should also be updated and ensured that they still work