Unable to ingest funders

[jeremyf]

British Library uses the crossref funder registry - each variation of the funder we've tried results in a failed import NoMethodError in Bx.

The following method is raising an exception, because the yielded hash is a String but the logic excpects a Hash.

britishlibrary/app/models/concerns/ubiquity/works_virtual_fields.rb

Lines 169 to 174 in 9c990a5

	def clean_incomplete_data_for_funder(data_hash)
	return if data_hash.blank?
	data_hash.each do |hash|
	hash.transform_values! { |_v| nil } if hash["funder_name"].blank?
	end
	end

The raw metadata regarding Funder is as follows:

funder_1: Arts and Humanities Research Council (AHRC)
funder_project_reference_1: AH/S01179X/1

The parsed metadata for Funder is as follows:

funder: ["[\"Arts and Humanities Research Council (AHRC)\"]"]

In the parsed metadata there is no instance of project reference.

Testing Instructions

altered sample file (Google Drive link)

• download the altered sample and create a Bulkrax csv importer with it
• check the funder of the resulting import
• observe that the funder appears on the show page
• observe that all the funder information is on the edit page

[j-basford]

See the examples:
https://bl-demo.iro.bl.uk/importers/97?locale=en
https://bl-demo.iro.bl.uk/importers/98?locale=en
Tried two variants of AHRC (the latter was Arts and Humanities Research Council which is what we see on the public view - perhaps we should have tried Arts and Humanities Research Council, United Kingdom, which is what you see in the lookup when selecting AHRC...)

[ShanaLMoore]

related to: #355

[ShanaLMoore]

@j-basford Would you happen to have the two files used to test this? The weird data issue on staging is making this hard to look into. Is this still a problem?

[ShanaLMoore]

@jeremyf Could you update this ticket with testing instructions or steps to reproduce?

[kirkkwang]

Looks like funder_project_reference_1 should really be fndr_project_ref_1 which is its own property and not a part of the funder hash object.

Here's what it looks like when I use fndr_project_ref_1 in the CSV:

The Funder Project Reference field does not show anywhere in the UI, but the value is stored and shows in the JSON feed.

Also, since Bulkrax does not have an API lookup for funder, the fields all need to be explicitly included in the CSV. The CSV would look like:

funder_name_1	funder_doi_1	funder_isni_1	funder_ror_1	funder_award_1
Arts and Humanities Research Council (AHRC)	http://dx.doi.org/10.13039/501100000267	0000 0004 3497 6001	https://ror.org/0505m1554

Here's the altered CSV for the failed example:
BLNewspapers_NuneatonTimes_test.csv

[laritakr]

Passed QA.

Show page:

Edit page:

[grahamjevon]

Passes BL QA - funder details successfully imported. Just need to be mindful that there is no API validation. So maybe we need to include a dropdown list in the bx template (to maintain consistency).