EML 2.2.0 testing - comparing EML from IPT to EML accessible from API

[ManonGros]

For this dataset https://www.gbif-uat.org/dataset/d2013cd8-ebeb-4320-90c3-d57469369d87
The API adds a role to the contact of the related project.
In the EML coming from the IPT we have:

            <relatedProject id="TEST2">
                <title>Hello</title>
                <personnel>
                    <individualName>
                        <surName>another test</surName>
                    </individualName>
                    <role></role>
                </personnel>
            </relatedProject>

While the EML accessible via the API shows

                    <relatedProject id="TEST2">
                        <title>Hello</title>
                        <personnel>
                            <individualName>
                                <surName>another test</surName>
                            </individualName>
                            <role>ADMINISTRATIVE_POINT_OF_CONTACT</role>
                        </personnel>
                    </relatedProject>

Another difference I noticed (and I think it isn't very important) is that the IPT puts country codes in the EML and the API spells out the country names. For example, <country>FR</country> in the IPT-generated EML and FRANCE in the API one.

[mike-podolskiy90]

Thanks Marie

I think it's a default behaviour. If an agent doesn't have a role it's ADMINISTRATIVE_POINT_OF_CONTACT

[mike-podolskiy90]

FR vs. FRANCE doesn't matter I think, both are valid values.

[mike-podolskiy90]

Dataset's EML is invalid though, I thought I added a validation when publishing. Odd.

Update: Yes, it's still possible to publish an invalid EML, but only for Metadata-only resources. Need to fix that.

[mike-podolskiy90]

This seems fine, closing

[mike-podolskiy90]

related #2473