fix Adaptor issues when HGTD is not present in DetectorDescription
Fixes issues seen in ART tests where GeoSiHit was being called in HitAnalysis/Tests - only try to initialize a HGTD_ID if HGTD has a DetectorManager present.
Also change behaviour of initialization - now it will only call init() if no DetectorManager is found, rather than if not all DetectorManagers are found. The former behaviour caused tonnes of WARNINGS if running without any subsytem present (which is a valid use case, for e.g. debugging pixels stand-alone). Since the Detector Description should not be changing within a job in terms of what subsystems are present, this should be safe.