Special case 10.5555 DOIs (and fix ambiguous DOI checker results)

[sneakers-the-rat]

There's a recurring issue where the DOI checker flags "10.5555" prefixed DOIs as missing with the ambiguous message I left in #101 (my bad). Those DOIs are known to not resolve.

This PR special-cases 10.5555 prefixed DOIs such that

• if they are present in the {doi} field they are marked as invalid with a prompt to include it in the {url} field instead
• If they are present in the {url} field they are marked as :skip with a note for the editor to confirm they resolve - I tried testing the resolution, but unhelpfully the website seems to respond with a 200 code and display a 404 page......

To do that, I

• moved the body of the checker to a handle_missing_doi method
• made a handle_special_case method to allow for future special cases, if any

I also fixed the ambiguous message by adding a skip type that is "maybe ok, editor take a look" field and made the template render categories with emoji that indicate the status of each DOI, it looks like...

Reference check summary (note 'MISSING' DOIs are suggestions that need verification):

✅ - OK DOIs

- this value is ok
- another ok value

❔ - SKIP DOIs

- this value is skipped but may still be ok
- https://dl.acm.org/doi/10.5555/xxxx.yyyy  - correctly put 10.5555 prefixed doi in the url field, editor should ensure this resolves

❌ - MISSING DOIs

- this one needs a doi
- so does this one

❌ - INVALID DOIs

- 10.5555/xxxx.yyyy is INVALID - 10.5555 is a known broken prefix, replace with https://dl.acm.org/doi/{doi} in the {url} field"

My ruby is bad, so someone else please check my work here!

[arfon]

Looks reasonable to me. @xuanxu – I think you usually take Buffy deploys out for a spin on Heroku – would you recommend we do a branch deploy here or are you happy to merge?

[xuanxu]

👍 Let's merge it

[xuanxu]

In action (I've changed the "skip" symbol to a more visible one):
openjournals/joss-reviews#7116 (comment)

[sneakers-the-rat]

yeah that looks better, i was looking for a yellow indicator with a question mark that wasn't the ⚠️ which seemed too threatening.

looking at that review^ it seems like maybe there could be more info for "no doi given and none found" kind of citations? i don't mean to overcomplicate this one very small part of the review system, it just seems to come up semi-frequently - i had meant to add a similar directive like "editor check if this resolves" for "editor manually check this reference" but if it works fine as-is then great we can call it shipped and move on <3.

[arfon]

no doi given and none found

I do think this language is a little confusing to people sometimes. @danielskatz had flagged this a while back too.

So yes, 👍 to making this more directive to editors/authors.

[danielskatz]

I suggested in Slack, and repeat here for tracking:

I wonder if we can change some DOI checker language? When I see this
https://dl.acm.org/doi/10.5555/1251460.1251470 - correctly put 10.5555 prefixed doi in the url field, editor should ensure this resolves
it's unclear to me what part of it is fact and what part is instructions, along with the error that this is not a DOI...

Could we instead say
https://dl.acm.org/doi/10.5555/1251460.1251470 - non-DOI with 10.5555 correctly placed in the url field, editor should ensure this resolves

And also maybe change the category for this, as Skip DOI doesn't make sense, since these aren't actually DOIs...