Update app.py

app.py

@@ -95,6 +95,37 @@ app.layout = dbc.Container([
     dcc.Store(id='generated-audio'),
 ])
 
+def process_prompt(text, voice, tokenizer, device):
+    prompt = f"{voice}: {text}"
+    inputs = tokenizer(prompt, return_tensors="pt")
+    input_ids = inputs["input_ids"].to(device)
+    attention_mask = inputs["attention_mask"].to(device)
+    return input_ids, attention_mask
+
+def parse_output(generated_ids):
+    decoded = tokenizer.decode(generated_ids[0], skip_special_tokens=True)
+    code_list = [int(code) for code in decoded.split() if code.isdigit()]
+    return code_list
+
+def redistribute_codes(code_list, snac_model):
+    audio = snac_model.codes_to_audio(torch.tensor(code_list).unsqueeze(0).to(device))
+    return audio.cpu().numpy().flatten()
+
+def detect_silence(audio, threshold=0.01, min_silence_len=1000):
+    is_silent = np.abs(audio) < threshold
+    silent_regions = []
+    silent_start = None
+    for i, silent in enumerate(is_silent):
+        if silent and silent_start is None:
+            silent_start = i
+        elif not silent and silent_start is not None:
+            if i - silent_start >= min_silence_len:
+                silent_regions.append((silent_start, i))
+            silent_start = None
+    if silent_start is not None and len(audio) - silent_start >= min_silence_len:
+        silent_regions.append((silent_start, len(audio)))
+    return silent_regions
+
 # Combined callback
 @callback(
     Output("script-output", "value"),
@@ -123,7 +154,6 @@ app.layout = dbc.Container([
     State("advanced-settings", "is_open"),
     prevent_initial_call=True
 )
-
 @spaces.GPU()
 def combined_callback(generate_script_clicks, generate_audio_clicks, advanced_settings_clicks, clear_clicks,
                       host1_name, host2_name, podcast_name, podcast_topic, prompt, uploaded_file, duration, num_hosts,