fixed ad filter to prevent recognizing all audios fingerprinted in box

ondemand/service.py

@@ -99,7 +99,8 @@ def process_segment(item):
     results = []
     for path, name, ts in iterate_audios(date, station):
         print(path)
-        for match in dejavu.recognize(recognizer, path, 5):
+        for match in dejavu.recognize(recognizer, path, 5,
+                                      ads_filter=[md5hash]):
             try:
                 results.append({
                     'confidence': match['confidence'],