EJP output validation

[Melissa37]

Problem / Motivation

WHO: Production/Production Vendor WHEN: Between Export from EJP/xPub and delivery to production vendor WHERE: In eLife bot processes WHAT: Validate what is provided against what is expected WHY: EJP regression testing is poor and when they make changes to their code it often affects the export and eLife does not know something has dropped off. It should not be the responsibility of the production vendor to validate this content, and as such we cannot blame then when something drops off and they don't tell us, or we don't find out for a few days. We'd like to ensure we identify issues as soon as they happen

Proposed solution

• A validation process occurs and if any file fails it is not passed to the production vendor but an alert sent to the production team
• Production still need to allow the content to be sent on to Production Vendor either with details of what's missing and what to do, or with the ability to correct the XML manifest or other files

Validate input as soon as the data is available in a bucket, and stop before sending output on.

How do you suggest the validation alerts to be received or shown - email? dashboard? a queue reader? Email initially.

Would it be satisfactory initially if you wanted to change some metadata it would require downloading the input zip file, modifying it locally on your computer, zipping it back up and adding it to the bucket again, with a slightly different file name? Yes, but why would it need a different name? PoA process does not do that?

If you are appending notes for the Production Vendor, how would you supply these? Do not think we'd be doing that

Clarification needed and assumptions

• We need to devise a process so production can still allow delivery of the files because it can take weeks for EJP to fix errors and we cannot halt production while that happens

---

Tasks

• [ ] Review list of expected output and values (other tickets will feed into this), for example: Each author must have an EJP ID Each author must have an affiliation Each author must have a contribution There has to be a funding statement Each file with @file-type='supp' provided has to have the following XML fields completed: <custom-meta><meta-name>Title</meta-name><meta-value>Supplementary File 3</meta-value></custom-meta></file> BUT we might want to change these metadata fields on conversion

Technical notes

@gnott

User interface / Wireframes

@gnott @eLifeProduction

[gnott]

We can validate input as soon as the data is available in a bucket, and we can stop before sending output on.

How do you suggest the validation alerts to be received or shown - email? dashboard? a queue reader?

Would it be satisfactory initially if you wanted to change some metadata it would require downloading the input zip file, modifying it locally on your computer, zipping it back up and adding it to the bucket again, with a slightly different file name?

If you are appending notes for the Production Vendor, how would you supply these?

[Melissa37]

Hi G, updated the main ticket.

Thanks! M