Updated phonopy processing so that the smaller subsystem is used as...

Updated phonopy processing so that the smaller subsystem is used as representative one and that calculation references are specified relative to the upload root.

phonopy @ 1ab5a1ee

-Subproject commit 09bc61058470381a9d590e70cf9ee0c76fb120f2
+Subproject commit 1ab5a1eee99b58451b9d818d1b0704b0005dc5c9

nomad/normalizing/normalizer.py

@@ -141,6 +141,18 @@ class SystemBasedNormalizer(Normalizer, metaclass=ABCMeta):
                     'chose "representative" system for normalization',
                 )
 
+        # If the found system is referencing a subsystem, then we choose it as
+        # the representative one. These smaller subsystems are much easier to
+        # analyze. Currently used in phonon calculations.
+        if system is not None:
+            try:
+                system_ref = system.section_system_to_system_refs[0]
+                ref_kind = system_ref.system_to_system_kind
+                if ref_kind == "subsystem":
+                    system = system_ref.system_to_system_ref
+            except Exception:
+                pass
+
         if scc is not None:
             self.section_run.m_cache["representative_scc_idx"] = scc.m_parent_index
         if system is not None:

nomad/processing/data.py

@@ -417,8 +417,7 @@ class Calc(Proc):
             # an absolute path which needs to be converted into a path that is
             # relative to upload root.
             scc = backend.entry_archive.section_run[0].section_single_configuration_calculation[0]
-            ref = scc.section_calculation_to_calculation_refs[0].calculation_to_calculation_external_url
-            relative_ref = ref.split("/", 6)[-1]
+            relative_ref = scc.section_calculation_to_calculation_refs[0].calculation_to_calculation_external_url
             ref_id = upload_files.calc_id(relative_ref)
             with upload_files.read_archive(ref_id) as archive:
                 arch = query_archive(archive, {ref_id: ref_id})[ref_id]