updated output format

app.py

@@ -1,4 +1,4 @@
-import json
+from pprint import pformat
 
 import numpy as np
 
@@ -19,6 +19,7 @@ LM_HUB_FP = 'language_model/cv8be_5gram.bin'
 
 def main(audio_fp: str):
     audio, sampling_rate = torchaudio.load(audio_fp, normalize=True)
+    init_audio_shape = audio.shape
 
     # convert stereo to mono
     converted_to_mono = False
@@ -39,16 +40,18 @@ def main(audio_fp: str):
 
     # recognize speech
     res = pipeline(inputs=inputs)
-    # text_recognized = res['text'][0]
 
+    # add additional information to the output
+    res['text'] = res['text'][0]  # unpack batch of size 1
     res['sampling_rate_orig'] = sampling_rate
-    res['init_audio_shape'] = audio.shape
+    res['init_audio_shape'] = init_audio_shape
     res['converted_to_mono'] = converted_to_mono
+    res['resampled_audio_shape'] = audio_resampled.shape
     res['inputs_shape'] = inputs.shape
     res['inputs_max'] = np.max(inputs).item()
     res['inputs_min'] = np.min(inputs).item()
 
-    res_str = json.dumps(res, indent=2)
+    res_str = pformat(res)
 
     return res_str