incorporating the PR comments, adding sanity check

- not more than 20% utterances can be dropped on `kaldi import`

lhotse/kaldi.py

@@ -148,13 +148,16 @@ def fix_id(t: str) -> str:
 
     # remove recordings with 'None' duration (i.e. there was a read error)
     for recording_id, dur_value in durations.items():
-        if dur_value == None:
+        if dur_value is None:
             logging.warning(
                 f"[{recording_id}] Could not get duration. "
                 f"Failed to read audio from `{recordings[recording_id]}`. "
-                f"Dropping the recording from manifest."
+                 "Dropping the recording from manifest."
             )
             del recordings[recording_id]
+    # make sure not too many utterances were dropped
+    if len(recordings) < len(durations) * 0.8:
+        raise RuntimeError(f"Failed to load more than 20% utterances of the dataset: \"{path}\"")
 
     # assemble the new RecordingSet
     recording_set = RecordingSet.from_recordings(