XML serialisation of instances with the same URLs #167

This PR integrates the changes needed to adapt the instance XML serialization proceduce as described in #7156

Instances with multiple URLs

We can choose one, because ARC told us that they are only displaying the first one anyway.
Decision : CNR can pick one URL for the instance based on one of the following strategies: 1. take the first, for simplicity we start with this approach. 2. take a URL which is a DOI, a Handle (or anything else we can recognise). if not available, fallback on option 1.

Instances with different URL that resolve to the same page
Decision: we cannot do anything about it. We live with it.

Instances with the same URL

After the previous step, each instance has one URL, but the URL can be the same among different instances.
Decision : merge all instances with the same URL into one.

Merged instances will feature

most open accessright
multiple collectedfrom (to be shown as comma separated in the Provider)
multiple hostedby - the unknown repository is ruled out when a valid repo exist
apc sum and currency will not be repeatable. If, for reasons we cannot imagine, two different APC information are attached to the same URL, then the last one will win.
all other fields: allow multiple unique values

This PR integrates the changes needed to adapt the instance XML serialization proceduce as described in [#7156](https://support.openaire.eu/issues/7156) **Instances with multiple URLs** * We can choose one, because ARC told us that they are only displaying the first one anyway. * Decision : CNR can pick one URL for the instance based on one of the following strategies: 1. take the first, for simplicity we start with this approach. 2. take a URL which is a DOI, a Handle (or anything else we can recognise). if not available, fallback on option 1. **Instances with different URL that resolve to the same page** **Decision**: we cannot do anything about it. We live with it. **Instances with the same URL** * After the previous step, each instance has one URL, but the URL can be the same among different instances. Decision : merge all instances with the same URL into one. Merged instances will feature * most open accessright * multiple collectedfrom (to be shown as comma separated in the Provider) * multiple hostedby - the unknown repository is ruled out when a valid repo exist * apc sum and currency will not be repeatable. If, for reasons we cannot imagine, two different APC information are attached to the same URL, then the last one will win. * all other fields: allow multiple unique values

Labels Milestones

XML serialisation of instances with the same URLs #167

Step 1:

Step 2: