Allow non-expanded datasetRefs in Butler.transfer_from.

Previously, transfer_from did not impose requirements on dataset refs,
but set a flag on _importDatasets that (per its docs) should only have
been set if the caller could guarantee expanded datasetRefs.

This commit keeps the open-ended precondition on transfer_from, and
allows internal expansion instead. This should be a no-op if the input
datasetRefs are already expanded. The expand flag is left at its
default instead of being explicitly set to True to give _importDatasets
ultimate control over expansion (e.g., in case it no longer requires
expanded IDs in the future).

doc/changes/DM-41165.bugfix.rst

@@ -0,0 +1,2 @@
+Butler.transfer_from no longer requires expanded dataset refs under certain circumstances.
+However, providing expanded refs in advance is still recommended for efficiency.

python/lsst/daf/butler/_butler.py

@@ -2296,7 +2296,8 @@ def transfer_from(
             `Butler` whose registry will be used to expand data IDs.
         source_refs : iterable of `DatasetRef`
             Datasets defined in the source butler that should be transferred to
-            this butler.
+            this butler. In most circumstances, ``transfer_from`` is faster if
+            the dataset refs are expanded.
         transfer : `str`, optional
             Transfer mode passed to `~lsst.daf.butler.Datastore.transfer_from`.
         skip_missing : `bool`
@@ -2466,7 +2467,7 @@ def transfer_from(
 
                 # Assume we are using UUIDs and the source refs will match
                 # those imported.
-                imported_refs = self._registry._importDatasets(refs_to_import, expand=False)
+                imported_refs = self._registry._importDatasets(refs_to_import)
                 assert set(imported_refs) == set(refs_to_import)
                 n_imported += len(imported_refs)