respect verbose flag

jasper/data/asr_recycler.py

@@ -59,10 +59,12 @@ def extract_data(
             code_wav = code_fb.getvalue()
             # only starting 1 min audio has reliable alignment ignore rest
             if start_time > 60:
+                if verbose:
                     print(f'start time over 60 seconds of audio skipping.')
                 break
             # only if some reasonable audio data is present yield it
             if code_seg.duration_seconds < 0.5:
+                if verbose:
                     print(f'transcript chunk "{transcript}" contains no audio skipping.')
                 continue
             yield transcript, code_seg.duration_seconds, code_wav

jasper/data/rev_recycler.py

@@ -63,6 +63,7 @@ def extract_data(
             # print(monologue["speaker_name"])
             speaker_channel = channel_map.get(monologue["speaker_name"])
             if not speaker_channel:
+                if verbose:
                     print(f'unknown speaker tag {monologue["speaker_name"]} in wav:{wav_path} skipping.')
                 continue
             try:
@@ -79,6 +80,7 @@ def extract_data(
                     .collect()(monologue)[-1]
                 )
             except IndexError:
+                if verbose:
                     print(f'error when loading timestamp events in wav:{wav_path} skipping.')
                 continue
 
@@ -92,6 +94,7 @@ def extract_data(
             text_clean = re.sub(r"\[.*\]", "", text)
             # only if some reasonable audio data is present yield it
             if tscript_wav_seg.duration_seconds < 0.5:
+                if verbose:
                     print(f'transcript chunk "{text_clean}" contains no audio in {wav_path} skipping.')
                 continue
             yield text_clean, tscript_wav_seg.duration_seconds, tscript_wav
@@ -113,6 +116,7 @@ def extract_data(
                     .collect()(monologue)[-1]
                 )
             except IndexError:
+                if verbose:
                     print(f'error when loading timestamp events in wav:{wav_path} skipping.')
                 continue
 
@@ -125,6 +129,7 @@ def extract_data(
             text = "".join(lens["elements"].Each()["value"].collect()(monologue))
             text_clean = re.sub(r"\[.*\]", "", text)
             if tscript_wav_seg.duration_seconds < 0.5:
+                if verbose:
                     print(f'transcript chunk "{text_clean}" contains no audio in {wav_path} skipping.')
                 continue
             yield text_clean, tscript_wav_seg.duration_seconds, tscript_wav