GitLab

[hook] Fix dispatching for file_category to format_id / mime_type on import of an archive unit component

by changing related hook's category to metadata, since others are deactivated
during sub-tree cloning.

overriden function now returns a generator, we've to turn it to a list.

consistently with other fields

Monkeypatches for derived rtags introduced a regression in SedaArchiveUnit
form. Since derived rtags are now relased in CubicWeb 3.25, let's move forward
and drop our broken monkeypatchs.

Related to extranet #37884173

Follow-up on cc3b3e0f6d42 to fix recurring "AttributeError: 'NoneType'
object has no attribute 'concepts'" observed in demo's logs.

causing error at import time :'(

which will occurs when the profile has some data object linked to several mime
types or format ids. Tests setup is shared with profile generation, so modify it
there, it won't hurt the original functional case to cover this as well.

Closes #36331831

since this may now occurs with values deducted from file category. In such case,
generates proper rng:choices / xsd:enumeration.

Related to #36331831

Upon modification of file_category relation, synchronize mime_type / format_id
values for the binary data-object, through intermediary SEDAMimeType /
SEDAFormatId entities which are kept for now but could be removed if the
automatic user cardinality system is validated by business people.

Synchronization is implemented to consider textual values of concept in the
category vocabulary, and to keep only those which have a match in the
vocabularies that are set on the transfer (code list version).

Because of that, don't attempt to synchronize values on archive unit component,
that should be done when it is imported into a transfer.

Related to #36331831

on a binary data object. This is done to prepare automatic synchronization of
these entities'values (concept) according to file_category. At some point we may
want to remove them in favor of direct relations to the binary data object,
since these are there to handle a cardinality which will be automatically
handled, provided the implemented behaviour fine with business people.

Related to #36331831

rather than later, so one may still use the `import_seda_schemes` high-level
function to import files which are not in the migration/data directory (but
rather in test/data).

Related to #36331831

Do it minimal with no js involved but a simple multiple select. This still have
to be coupled to format_id / mime_type but this will be done in later csets.

Related to #36331831

So they may link to several mime types / formats deducted from the file
category. This was actually not very usable before, since one could not expect
a single pronom format (e.g. for a specific PDF version).

Related to #36331831

provided by the SAEM team. Link it to the file_category relation.

Related to #36331831

This relation will be used in later csets to link the file to a high-level file
format, that will be used to control low-level mime  type and format id fields.

Related to extranet #36331831

not  format_litteral

in case the profile is a simplified profile. This is because this relation is
there because of the SEDA 2 underlying model, where all data objects are below
the transfer node, but it is not used in SEDA 0.2 and 1.0, where data
objects are dispatched under the archive units tree. So we don't want to prevent
having several data objects with cardinality != 1 at different level of the
tree.

Closes extranet #37372358

Introduce a bootstrap tooltip whose presence is indicated through a question
mark on hover.

Notice the comment change above aliasing cw's uicfg: this have to be there
because of the relative import from our own uicfg module, though it's still
unclear to me why this screw up the module's local namespace.

Closes #17106808

Introduce in views/uicfg.py a new data structure providing XSD element name and
documentation indexed by entity type, relation type and role.

Related to #17106808