GitLab

* hide all SEDA fields but archival agreement from primary view,

* hide all SEDA fields from main form, since we don't want to handle the
  simplified form switch yet,

* hide code list versions, data objects and related transfers tabs.

Display data objects as children of archive units, not as children of the
transfer.

Related to #14066262

* generate buttons to add binary/physical data object instead of reference, with
  extra information allowing to create automatically the reference

* override data object autoform's `inlined_form_views` method to insert the
  aforementioned information as an hidden field (this won't work with __linkto
  since this is a relation to an inlined entity, not the main entity)

* similar display of exising data objects as in the original data objects tab of
  archive transfer

Related to #14066262

This completes the previous cset which add basic rtags, by including a single
form instead of 'add' links. This form has no visible fields so nothing is
displayed with the help of the proper inlined form renderer.

Related to #14066262

So that a reference is created transparently at the same time. This is because
in simplified mode, we want to see data objects as children of archive units,
not as first class object which are later referenced by archive units.

Two distincts rules are added for SEDADataObjectReference and
SEDACustodialHistoryFile else it won't work because of the ambiguity.

Everything is then hidden in the SEDADataObjectReference form. In this mode, we
should eventually ignore its cardinality and only consider the one of the data
object.

Related to #14066262

To be reused from other contexts.

Related to #14066262

+ related tests, including tests for features provided by underlying's
`parent_and_container` function.

Related to #14066262

So underlying things such as deepcopy are hidden a bit. Along the way, rename dor_afs to
do_ref_afs which is easier to read.

We'll need that here soon!

This has two effects:
* better usage of horizontal space
* avoid the edition button to be misplaced

The second point was the original problem I was attempting to solve. We'll need
better if it appears that the new line introduced by the div was actually better
for readability.

Because SEDAContent had two list item views returning the same score.

This will be used in later csets to have a much limited, hence simpler, UI. This
UI should be similar as the one we have in SAEM-Ref (aka the "convergence" task).

Related to #14066262

Closes #11447493

we want two distinct actions for XSD and RNG export, hence two different regid.

Closes #11447493

In this case, return the 'intermediary' entity if any, it appears to be the end
of the path anyway.

This avoid allowing undefined elements in the profile.

Following the policy "no entity, not allowed", we should not allow jumped
elements when those have no content.

This was allowing 'FileInfo' accidentally in the functional test, as well as Gps
intermediary tag. Test and fix this.

Remove print and drop XXX in favor of explanation.

instead of a pyxst attribute object. This is preparation work for a later cset.

Note that each attribute may also have its own annotation, which are not
displayed in the list item view. IMO, we should remove those at some point
rather than cluttering this view.

What triggered this is still unclear though.

with nicer and consistent ordering w/ autoform.