In order to clean documents which are obsolete we need to be able to find those documents in both DCP1 and DCP2. For DCP1 we should be able to use search API but in DCP2 it's constrained so we'll have to improve developer_materials query to handle this additional case.