json-cache whisper transcription

scripts/transcribe-voice-tracks.py

@@ -4,6 +4,8 @@ usage = 'python transcribe-voice-track.py <fountain filename> <character> <wav f
 
 import util
 from imports import *
+import json
+from os.path import exists
 
 import whisper
 model = whisper.load_model("turbo")
@@ -80,7 +82,15 @@ map = fmap.map
 print(map)
 
 for audio_filename in audio_filenames:
-    result = model.transcribe(audio_filename)
+    result = None
+    if exists(f'{audio_filename}.json'):
+        with open(f'{audio_filename}.json', 'r') as f:
+            result = json.load(f)
+    else:
+        result = model.transcribe(audio_filename)
+        with open(f'{audio_filename}.json', 'w') as f:
+            json.dump(result, f)
+
     print(result['segments'])
     for segment in result['segments']:
         match = fmap.best_match(segment['text'])