Invalid DOI?

[JessyBarrette]

I'm not sure if this is related to lates the changes not been include on the form yet. But here's one metadata record that can't be submitted for review due to an invalid DOI

https://cioos-siooc.github.io/metadata-entry-form/#/en/hakai/7U7b8oPpeTN6gjvXlUCTGJr5pga2/-McQFPAf457LB4-SWmyL

@Br-Johnson @timvdstap @sorochak any thoughts?

[timvdstap]

Was that DOI generated through the form? If it was manually put in, try adding https://doi.org/ in front of it.

See e.g. https://github.com/cioos-siooc/metadata-entry-form/issues/230

[JessyBarrette]

Ok I thought it was generated by the form but potentially it was generated manually. Likely when those others were generated:

https://github.com/HakaiInstitute/hakai-datasets/issues/70

[JessyBarrette]

This still prevents us from publish this dataset. I have no control on the DOI myself. And I believe that DOI is still in Draft

This leads to nothing: https://doi.org/10.21966/fg51-ms50

[timvdstap]

I mean I guess it makes sense that it doesn't lead to anything yet given that the record/DOI is in draft. But that shouldn't be preventing the record to being published... Have we tried deleting the draft DOI and generating a new one? I'm not sure how desirable that is given that this record has been around for some time and the DOI may have already been shared.

[JessyBarrette]

yeah that old Draft was never used, I can delete and regenerate a new one, which I just did. DOI is now 10.21966/90h6-b497

Now, the Draft DOI was generated by the form. However, I can't still submit the form due to an Invalid DOI 😞

[timvdstap]

@n-a-t-e do you perhaps have any thoughts on this?

[sorochak]

The invalid DOI message occurs because our DOI validation checks to make sure it is a valid Http URL.

The simplest solution is to remove the http check and remove http://doi/ from any current records that have it. The datacite api doesn't handle urls either, so we would have to add/remove the url portion on create/update.

What was the reasoning behind adding the http url portion to the DOIs ?

@JessyBarrette @n-a-t-e @Br-Johnson @timvdstap @fostermh

[n-a-t-e]

it can be stored with or without that http://doi/ . Either way, all the records in the DB need to be updated to be either with or without http://doi/ for consistency

One reason to store it with http in the database, is there are at least 3 places I can see where this DOI gets used, in its full URL form:

• EML
• ISO XML
• erddap XML snippet

So if we don't store with http://doi/ in the DB, we just have to tack it on in those 3 spots, not hard

Also, how we store in the DB vs what we ask on the form dont need to be related, we can store full URL and just ask for 10.000/1234 in the form, and then split/reform as needed

[fostermh]

I am in favour of storing it as the full url. it is easy to strip off the https://doi.org/ part when making calls to datacite and it avoids ambiguity in down stream applications that use the identifier.

[sorochak]

@JessyBarrette the doi creation in the form now prepends the URL to the generated datacite DOI. This should fix your issue.

[sorochak]

All DOIs in the DB are now be stored as the full url.

[JessyBarrette]

Awesome thanks everyone the record associated with this issue also succeeded to be submitted. I think I can now close this issue.