Update app.py

app.py

@@ -1,7 +1,6 @@
 import gradio as gr
-import requests
-import random
-import urllib.parse
+import edge_tts
+import asyncio
 import tempfile
 import os
 import nltk
@@ -9,125 +8,98 @@ import srt
 from pydub import AudioSegment, silence
 import datetime
 
-nltk.download("punkt")
+nltk.download('punkt')
 
-NSFW_URL_TEMPLATE = os.getenv("NSFW_API_URL_TEMPLATE")
-TTS_URL_TEMPLATE = os.getenv("TTS_API_URL_TEMPLATE")
+# 🔊 Generate TTS audio
+async def text_to_speech(text, voice, rate, pitch):
+    if not text.strip():
+        return None, None, "Please enter text to convert."
+    if not voice:
+        return None, None, "Please select a voice."
 
-if not NSFW_URL_TEMPLATE:
-    raise ValueError("Missing Secret: NSFW_API_URL_TEMPLATE is not set.")
-if not TTS_URL_TEMPLATE:
-    raise ValueError("Missing Secret: TTS_API_URL_TEMPLATE is not set.")
+    voice_short_name = voice.split(" - ")[0]
+    rate_str = f"{rate:+d}%"
+    pitch_str = f"{pitch:+d}Hz"
+    communicate = edge_tts.Communicate(text, voice_short_name, rate=rate_str, pitch=pitch_str)
 
-VOICES = ["alloy", "echo", "fable", "onyx", "nova", "shimmer", "coral", "verse", "ballad", "ash", "sage", "amuch", "dan"]
+    with tempfile.NamedTemporaryFile(delete=False, suffix=".mp3") as tmp_file:
+        tmp_path = tmp_file.name
+        await communicate.save(tmp_path)
 
-def generate_srt(audio_path, script_text):
+    # Generate SRT
+    srt_path = generate_srt(tmp_path, text)
+
+    return tmp_path, srt_path, None
+
+# 🧠 Generate SRT from audio + text
+def generate_srt(audio_path, text):
     audio = AudioSegment.from_file(audio_path)
-    silences = silence.detect_silence(audio, min_silence_len=300, silence_thresh=audio.dBFS - 16)
-    silences = [(start / 1000.0, end / 1000.0) for start, end in silences]
-    sentences = nltk.tokenize.sent_tokenize(script_text)
+    silence_ranges = silence.detect_silence(audio, min_silence_len=400, silence_thresh=audio.dBFS - 16)
+    silence_ranges = [(start / 1000.0, end / 1000.0) for start, end in silence_ranges]
+    sentences = nltk.tokenize.sent_tokenize(text)
 
     subtitles = []
     last_time = 0.0
     for i, sentence in enumerate(sentences):
-        if i < len(silences):
+        if i < len(silence_ranges):
             start = last_time
-            end = silences[i][0]
-            last_time = silences[i][1]
+            end = silence_ranges[i][0]
+            last_time = silence_ranges[i][1]
         else:
             start = last_time
-            end = start + 2.5  # default
-        subtitle = srt.Subtitle(
+            end = start + 2.5  # fallback timing
+        subtitles.append(srt.Subtitle(
             index=i + 1,
             start=datetime.timedelta(seconds=start),
             end=datetime.timedelta(seconds=end),
             content=sentence
-        )
-        subtitles.append(subtitle)
+        ))
 
     srt_data = srt.compose(subtitles)
     with tempfile.NamedTemporaryFile(delete=False, suffix=".srt", mode='w') as srt_file:
         srt_file.write(srt_data)
         return srt_file.name
 
-def check_nsfw(prompt: str) -> bool:
-    encoded_prompt = urllib.parse.quote(prompt)
-    url = NSFW_URL_TEMPLATE.format(prompt=encoded_prompt)
-    try:
-        response = requests.get(url, timeout=20)
-        result = response.text.strip().upper()
-        return result == "YES"
-    except:
-        raise gr.Error("Failed to check prompt safety.")
-
-def generate_audio(prompt: str, voice: str, emotion: str, seed: int) -> bytes:
-    encoded_prompt = urllib.parse.quote(prompt)
-    encoded_emotion = urllib.parse.quote(emotion)
-    url = TTS_URL_TEMPLATE.format(prompt=encoded_prompt, emotion=encoded_emotion, voice=voice, seed=seed)
-    response = requests.get(url, timeout=60)
-    if 'audio' not in response.headers.get('content-type', ''):
-        raise gr.Error("Invalid audio response.")
-    return response.content
-
-def text_to_speech_app(prompt, voice, emotion, use_random_seed, specific_seed, subtitle_script):
-    if not prompt or not voice:
-        raise gr.Error("Prompt and Voice are required.")
-    seed = random.randint(0, 2**32 - 1) if use_random_seed else int(specific_seed)
-    is_nsfw = False  # You can enable this by: is_nsfw = check_nsfw(prompt)
-
-    if is_nsfw:
-        return None, None, "Prompt is flagged NSFW"
-
-    try:
-        audio_bytes = generate_audio(prompt, voice, emotion, seed)
-        with tempfile.NamedTemporaryFile(delete=False, suffix=".mp3") as temp_audio:
-            temp_audio.write(audio_bytes)
-            audio_path = temp_audio.name
-
-        srt_path = None
-        if subtitle_script.strip():
-            srt_path = generate_srt(audio_path, subtitle_script)
-
-        return audio_path, srt_path, f"Audio and SRT generated with seed {seed}"
-    except Exception as e:
-        return None, None, f"Error: {str(e)}"
-
-def toggle_seed_input(use_random_seed):
-    return gr.update(visible=not use_random_seed, value=12345)
-
-with gr.Blocks() as app:
-    gr.Markdown("## 🎙️ Advanced OpenAI TTS + Subtitle Generator")
-
-    with gr.Row():
-        with gr.Column(scale=2):
-            prompt_input = gr.Textbox(label="Prompt", placeholder="Enter your text...")
-            emotion_input = gr.Textbox(label="Emotion Style", placeholder="happy, sad, excited, calm...")
-            voice_dropdown = gr.Dropdown(label="Voice", choices=VOICES, value="alloy")
-            subtitle_script = gr.Textbox(label="Subtitle Script", lines=6, placeholder="Paste script here for SRT generation")
-        with gr.Column(scale=1):
-            random_seed_checkbox = gr.Checkbox(label="Use Random Seed", value=True)
-            seed_input = gr.Number(label="Specific Seed", value=12345, visible=False, precision=0)
-
-    submit_button = gr.Button("🎧 Generate Audio + Subtitles", variant="primary")
-
-    with gr.Row():
-        audio_output = gr.Audio(label="Generated Audio", type="filepath")
-        srt_output = gr.File(label="Download SRT File")
-        status_output = gr.Textbox(label="Status")
-
-    random_seed_checkbox.change(
-        fn=toggle_seed_input,
-        inputs=[random_seed_checkbox],
-        outputs=[seed_input]
-    )
-
-    submit_button.click(
-        fn=text_to_speech_app,
-        inputs=[prompt_input, voice_dropdown, emotion_input, random_seed_checkbox, seed_input, subtitle_script],
-        outputs=[audio_output, srt_output, status_output],
-        concurrency_limit=30
-    )
+# 🎛️ Interface wrapper
+async def tts_interface(text, voice, rate, pitch):
+    audio, srt_file, warning = await text_to_speech(text, voice, rate, pitch)
+    if warning:
+        return None, None, gr.Warning(warning)
+    return audio, srt_file, None
+
+# 📋 Setup Gradio UI
+async def create_demo():
+    voices = await edge_tts.list_voices()
+    voice_dict = {f"{v['ShortName']} - {v['Locale']} ({v['Gender']})": v['ShortName'] for v in voices}
+
+    with gr.Blocks() as demo:
+        gr.Markdown("# 🎙️ Edge TTS + Subtitle Generator (.srt)")
+
+        with gr.Row():
+            with gr.Column():
+                text_input = gr.Textbox(label="Input Text", lines=5, placeholder="Enter your script here...")
+                voice_dropdown = gr.Dropdown(choices=[""] + list(voice_dict.keys()), label="Select Voice", value="")
+                rate_slider = gr.Slider(minimum=-50, maximum=50, value=0, label="Speech Rate (%)")
+                pitch_slider = gr.Slider(minimum=-20, maximum=20, value=0, label="Pitch (Hz)")
+                generate_btn = gr.Button("🎧 Generate Audio + SRT")
+
+            with gr.Column():
+                audio_output = gr.Audio(label="Generated Audio", type="filepath")
+                srt_output = gr.File(label="Download .srt Subtitle")
+                warning_output = gr.Markdown(visible=False)
+
+        generate_btn.click(
+            fn=tts_interface,
+            inputs=[text_input, voice_dropdown, rate_slider, pitch_slider],
+            outputs=[audio_output, srt_output, warning_output]
+        )
+
+    return demo
+
+async def main():
+    demo = await create_demo()
+    demo.queue(concurrency_count=10)
+    demo.launch(show_api=False)
 
 if __name__ == "__main__":
-    if NSFW_URL_TEMPLATE and TTS_URL_TEMPLATE:
-        app.launch()
+    asyncio.run(main())