Could you give examples of regular queries which you tried? I'd take a look to find whatever that is a ranking problem or the DHT network just lacks good results for it.
About first problem though.. I already trained small model to catch spam pages and push them at the bottom of results, I'd spend more time to train it to catch pages about drugs too.
About first problem though.. I already trained small model to catch spam pages and push them at the bottom of results, I'd spend more time to train it to catch pages about drugs too.