chore(hybridcloud) Add non-tuple storage to RpcPrimaryKeyMap (#71439)

In order to have RPC validation apply to all methods we need to not have
any tuples in method parameters/returns. Replace the tuple inside
RpcPrimaryKey with structured data. This will result in some redundant
data for a day or so until the tuples are removed.

Refs HC-1190

src/sentry/services/hybrid_cloud/import_export/model.py

@@ -52,6 +52,12 @@ def into_rpc(cls, base_filter: Filter) -> "RpcFilter":
         )
 
 
+class RpcPrimaryKeyEntry(RpcModel):
+    new_id: int
+    kind: ImportKind
+    slug: str | None
+
+
 class RpcPrimaryKeyMap(RpcModel):
     """
     Shadows `sentry.backup.dependencies.PrimaryKeyMap` for the purpose of passing it over an RPC
@@ -62,17 +68,38 @@ class RpcPrimaryKeyMap(RpcModel):
     """
 
     # Pydantic duplicates global default models on a per-instance basis, so using `{}` here is safe.
+    # Deprecated, use map_entries going forward as that property does not contain tuples.
     mapping: dict[str, dict[int, tuple[int, ImportKind, str | None]]] = {}
 
+    # Pydantic duplicates global default models on a per-instance basis, so using `{}` here is safe.
+    map_entries: dict[str, dict[int, RpcPrimaryKeyEntry]] = {}
+
     def from_rpc(self) -> PrimaryKeyMap:
         pk_map = PrimaryKeyMap()
-        pk_map.mapping = defaultdict(dict, self.mapping)
+        if self.map_entries:
+            pk_mapping: dict[str, dict[int, tuple[int, ImportKind, str | None]]] = defaultdict(dict)
+            for model_name, entries in self.map_entries.items():
+                entries_data = dict()
+                for old_id, entry in entries.items():
+                    entries_data[old_id] = (entry.new_id, entry.kind, entry.slug)
+                pk_mapping[model_name] = entries_data
+            pk_map.mapping = pk_mapping
+        else:
+            pk_map.mapping = defaultdict(dict, self.mapping)
         return pk_map
 
     @classmethod
     def into_rpc(cls, base_map: PrimaryKeyMap) -> "RpcPrimaryKeyMap":
         converted = cls()
         converted.mapping = dict(base_map.mapping)
+
+        mapping_entries = dict()
+        for model_name, entries in base_map.mapping.items():
+            mapping_entries[model_name] = {
+                old_id: RpcPrimaryKeyEntry(new_id=entry[0], kind=entry[1], slug=entry[2])
+                for old_id, entry in entries.items()
+            }
+        converted.map_entries = mapping_entries
         return converted