Former committees restypes

cds_migrator_kit/rdm/records/transform/models/research_committee.py

@@ -95,11 +95,10 @@ class ResearchCommitteeModel(CdsOverdo):
 
     _default_fields = {
         "custom_fields": {},
-        "resource_type": {"id": "publication-other"},
     }
 
 
 research_comm_model = ResearchCommitteeModel(
     bases=(rdm_base_publication_model,),
-    entry_point_group="cds_migrator_kit.migrator.rdm.rules.publication",
+    entry_point_group="cds_migrator_kit.migrator.rules.research_committee",
 )

cds_migrator_kit/rdm/records/transform/transform.py

@@ -422,6 +422,11 @@ def creators(json, key="creators"):
             return _creators
 
         def _resource_type(entry):
+            # `_resource_type_rank` is bookkeeping for the 980__/697C_
+            # resource_type rule and research_committee.py's report-number
+            # detection (see research.py:resource_type) - drop it before it
+            # reaches the final record.
+            entry.pop("_resource_type_rank", None)
             try:
                 return entry["resource_type"]
             except KeyError:
@@ -617,6 +622,8 @@ def field_departments(record_json, custom_fields_dict):
                 if result and result not in custom_fields_dict["cern:departments"]:
                     custom_fields_dict["cern:departments"].append(result)
                 elif not result:
+                    if department.lower() == "cern?":
+                        continue
                     subj = json_output["metadata"].get("subjects", [])
                     subj.append({"subject": department})
                     json_output["metadata"]["subjects"] = subj

cds_migrator_kit/rdm/records/transform/xml_processing/rules/base.py

@@ -551,8 +551,7 @@ def corporate_author(self, key, value):
         departments = self.get("custom_fields", {}).get("cern:departments", [])
         if department and department not in departments:
             departments.append(department)
-        self["custom_fields"]["cern:departments"] = departments
-        raise IgnoreKey("contributors")
+            self["custom_fields"]["cern:departments"] = departments
     if "b" in value:
         unit = value.get("b")
         if unit:
@@ -877,6 +876,14 @@ def related_identifiers_787(self, key, value):
             "relation_type": {"id": "references"},
             "resource_type": {"id": "publication-report"},
         },
+        "addendum to": {
+            "relation_type": {"id": "issupplementto"},
+            "resource_type": {"id": "publication-report"},
+        },
+        "complemented by": {
+            "relation_type": {"id": "issuplementedby"},
+            "resource_type": {"id": "publication-report"},
+        },
         "preprint": {
             "relation_type": {"id": "references"},
             "resource_type": {"id": "publication-preprint"},

cds_migrator_kit/rdm/records/transform/xml_processing/rules/research.py

@@ -798,20 +798,36 @@ def resource_type(self, key, value):
             subjects = self.get("subjects")
             subjects.append({"subject": value_a if value_a else value_b})
             self["subjects"] = subjects
-            raise IgnoreKey("resource_type")
-        raise UnexpectedValue(
-            "Unknown resource type (Publications)", value=best_value, field=key
-        )
-
-    if current:
-        current_key = next((k for k, v in mapping.items() if v == current), None)
-        current_rank = priority.get(current_key, float("inf"))
+        # This function is invoked once per repeated 980__/697C_ occurrence
+        # (see CdsOverdo.do), with `self["resource_type"]` accumulating the
+        # best match across calls. An unmapped value here doesn't mean the
+        # record has no resource type - another occurrence, processed later,
+        # may still resolve to a known one, so don't abort the whole record.
+        # If no occurrence ever resolves, the missing resource_type is
+        # caught downstream (see `_resource_type` in transform.py).
+        raise IgnoreKey("resource_type")
 
+    # `current` may be `resource_type`'s value from an earlier 980__/697C_
+    # occurrence on this same record, but it may also be a default seeded by
+    # the model's `_default_fields` (e.g. ResearchCommitteeModel sets
+    # resource_type to publication-other upfront) before this rule ever ran.
+    # Reverse-looking up `current` in `mapping` can't tell those apart - a
+    # seeded default happens to equal "alephdraft"'s mapped value, so it
+    # would get treated as an already-decided, low-priority match and block
+    # any later occurrence whose value isn't in `priority` (rank stays
+    # `inf`, and `inf < inf` is False). Track the rank of our own previous
+    # decision explicitly instead, so a seeded default never poisons this
+    # comparison.
+    current_rank = self.get("_resource_type_rank")
+
+    if current and current_rank is not None:
         if rank < current_rank:
+            self["_resource_type_rank"] = rank
             return mapping[best_value]
         else:
             raise IgnoreKey("resource_type")
     else:
+        self["_resource_type_rank"] = rank
         return mapping[best_value]