Update app.py

app.py

@@ -1,19 +1,57 @@
 import gradio as gr
 from TTS.api import TTS
+import time
+import os
 
-# Load TTS model (first time might take 1–2 minutes)
-tts = TTS(model_name="tts_models/en/ljspeech/tacotron2-DDC", progress_bar=True, gpu=False)
+# Available voice models
+models = {
+    "fast_pitch": "tts_models/en/ljspeech/fast_pitch",
+    "tacotron2": "tts_models/en/ljspeech/tacotron2-DDC",
+    "glow_tts": "tts_models/en/ljspeech/glow-tts"
+}
+
+# Current model loaded (default)
+current_model_key = "fast_pitch"
+tts = TTS(models[current_model_key], gpu=False)
+
+# Synthesize function
+def synthesize(text, selected_model=None):
+    global tts, current_model_key
+
+    # Switch model if needed
+    if selected_model and selected_model != current_model_key:
+        current_model_key = selected_model
+        tts = TTS(models[current_model_key], gpu=False)
 
-def synthesize(text):
     output_path = "output.wav"
+    start_time = time.time()
     tts.tts_to_file(text=text, file_path=output_path)
-    return output_path
+    total_time = time.time() - start_time
+
+    # Calculate RTF (approximate)
+    audio_duration = len(text.split()) / 2.5  # est. 2.5 words/sec
+    rtf = round(total_time / audio_duration, 3)
 
-# Gradio UI
-gr.Interface(
+    return output_path, {
+        "processing_time_sec": round(total_time, 3),
+        "real_time_factor": rtf,
+        "model_used": current_model_key
+    }
+
+# Gradio API
+api = gr.Interface(
     fn=synthesize,
-    inputs=gr.Textbox(label="Enter text to synthesize"),
-    outputs=gr.Audio(type="filepath", label="Generated Audio"),
-    title="🗣️ Coqui TTS - CPU Demo",
-    description="Enter text and generate speech using Coqui's Tacotron2-DDC model (free and fast).",
-).launch(share=True)
+    inputs=[
+        gr.Textbox(label="Input Text"),
+        gr.Dropdown(list(models.keys()), value="fast_pitch", label="Select Voice Model"),
+    ],
+    outputs=[
+        gr.Audio(type="filepath", label="Synthesized Audio"),
+        gr.JSON(label="Meta Info (Time, Model, RTF)")
+    ],
+    title="🗣️ TTS API with Model Selector",
+    description="Send text to convert it to speech and get metadata via Gradio API. Change the voice model dynamically."
+)
+
+api.launch()
+