Optimize SQL queries when searching a rule

Searching all rules then filtering in python the parent path is
more efficient than finding all the parent locations and finding
the matching rules.

stock_reserve_rule/models/stock_reserve_rule.py

@@ -82,7 +82,14 @@ def _constraint_fallback_location_id(self):
                     raise ValidationError(msg)
 
     def _rules_for_location(self, location):
-        return self.search([("location_id", "parent_of", location.id)])
+        # We'll typically have a handful of rules, so reading all of them then
+        # checking if they are a parent location of the location is pretty
+        # fast. Searching all the parent locations then the rules matching them
+        # can be much slower if we have many locations.
+        rules = self.search([]).filtered(
+            lambda rule: rule.location_id.parent_path.startswith(location.parent_path)
+        )
+        return rules
 
     def _eval_rule_domain(self, move, domain):
         move_domain = [("id", "=", move.id)]