Use raiseload in query loading options

Use [raiseload](https://docs.sqlalchemy.org/en/14/orm/loading_relationships.html#sqlalchemy.orm.raiseload) loading technique in query options
to raise an error on accessing attributes that haven't been eagerly
loaded. There have been suboptimal SQL queries in sir previously due to
missing columns in extrapaths or using hybrid attributes in fields or
extrapaths which caused lazy loading. Since the attributes we access in
SIR are fixed and remain the same, we should aim for loading all of those
eagerly. raiseload will raise an error with a stacktrace to help us fix
the issue if a lazy load happens, helping in performance debugging.

sir/schema/searchentities.py

@@ -9,7 +9,7 @@
     from xml.etree.cElementTree import tostring
 except ImportError:
     from xml.etree.ElementTree import tostring
-from sqlalchemy.orm import class_mapper, Load
+from sqlalchemy.orm import class_mapper, Load, raiseload
 from sqlalchemy.orm.attributes import InstrumentedAttribute
 from sqlalchemy.orm.descriptor_props import CompositeProperty
 from sqlalchemy.orm.interfaces import ONETOMANY, MANYTOONE
@@ -229,7 +229,7 @@ def build_entity_query(self):
                         load = defer_everything_but(class_mapper(model),
                                                     load,
                                                     *required_columns)
-                query = query.options(load)
+                query = query.options(load, raiseload("*"))
         if self.extraquery is not None:
             query = self.extraquery(query)
         return query