content_classifier.cod

module content_classifier

def classify_content
  intent "classify a message as safe, unsafe, or uncertain based on keyword signals"
  sig    (message: String) -> Label
  effects {}
  cand
    intent "unsafe — contains harmful keyword"
    when   or(
             contains(lower(message), "spam"),
             contains(lower(message), "hate"),
             contains(lower(message), "violence")
           )
    belief("unsafe", 0.90)
  cand
    intent "safe — contains approval keyword"
    when   or(
             contains(lower(message), "approved"),
             contains(lower(message), "verified")
           )
    belief("safe", 0.90)
  cand
    intent "uncertain — no keyword match"
    belief("uncertain", 0.75)

def main
  intent "test classify_content with a sample message"
  sig    () -> Label
  effects {}
  cand
    classify_content("This message contains spam")