Update app.py

app.py

@@ -3,7 +3,7 @@ from kokoro import KPipeline
 import soundfile as sf
 import io
 import os
-import time
+from langdetect import detect  # Language detection library
 
 # Install espeak-ng if not installed
 if not os.system("which espeak-ng"):
@@ -13,9 +13,9 @@ else:
     st.text("Installing espeak-ng...")
 
 # Streamlit App UI Setup
-st.title("Interactive Text-to-Speech with Kokoro")
+st.title("Text-to-Speech with Kokoro")
 
-# Expander section for language samples
+# Expander section to display information in multiple languages
 with st.expander("Sample Prompt!"):
     st.markdown("""
     - My name is Shukdev. (In English)
@@ -27,84 +27,125 @@ with st.expander("Sample Prompt!"):
     - 我叫苏赫德夫。(In Chinese)
     - 私の名前はスクデフです。(In Japanese)
     """)
-    
-# Sidebar Instructions and Configuration
+
 st.sidebar.header("Configuration & Instructions")
 
+# Sidebar Instructions
 st.sidebar.markdown("""
 ### How to Use the Text-to-Speech App:
-1. **Enter Text**: Input any text that you want to convert to speech.
-2. **Select Language**: Choose the language of the text.
-3. **Select Voice**: Choose the voice style.
-4. **Select Speed**: Use the slider to change the speech speed.
-5. **Add Background Music**: Optional background music for the speech (choose genre).
-6. **Generate Speech**: Click on **Generate Audio** to create speech.
-7. **Download**: Download the generated audio file.
+1. **Enter Text**: In the main text area, input any text that you want the model to convert to speech.
+   
+2. **Select Language**: 
+   - Choose the language of the text you are entering. Available options include:
+     - 🇺🇸 American English (`a`)
+     - 🇬🇧 British English (`b`)
+     - 🇪🇸 Spanish (`e`)
+     - 🇫🇷 French (`f`)
+     - 🇮🇳 Hindi (`h`)
+     - 🇮🇹 Italian (`i`)
+     - 🇧🇷 Brazilian Portuguese (`p`)
+     - 🇨🇳 Mandarin Chinese (`z`)
+     - 🇯🇵 Japanese (`j`)
+   
+3. **Select Voice**:
+   - Choose the voice style for the speech. You can pick different voices based on tone and gender, such as `af_heart`, `af_joy`, etc.
+   
+4. **Adjust Speed**:
+   - Use the speed slider to change how fast the speech is generated. You can set it between `0.5x` to `2.0x`, where `1.0x` is the normal speed.
+5. **Generate Speech**:
+   - After configuring the settings, click on the **"Generate Audio"** button. The app will process your text and produce speech audio accordingly.
+   
+6. **Download**:
+   - Once the audio is generated, you can play it directly in the app or download it as a `.wav` file by clicking on the **"Download Audio"** button.
+Enjoy experimenting with the text-to-speech conversion, and feel free to try different voices, speeds, and languages!
 """)
 
+st.sidebar.markdown("""
+        ### Courtesy: [Kokoro](https://huggingface.co/hexgrad/Kokoro-82M?fbclid=IwY2xjawIKqzxleHRuA2FlbQIxMAABHaf9GldgYOzXktNuoRtNKqd-aL7r-S7zPGyC8ttYOiG2zYfQqLyV4Qm75A_aem_0wKLC2C87ZZ2F04WjPJbtA)
+    """)
+
+# Language Detection Function
+def detect_language(text):
+    try:
+        lang = detect(text)
+        return lang
+    except Exception as e:
+        st.error("Error detecting language: " + str(e))
+        return None
+
 # User input for text, language, and voice settings
 input_text = st.text_area("Enter your text here", placeholder="The sky above the port was the color of television...")
-lang_code = st.selectbox("Select Language", ['a', 'b', 'e', 'f', 'h', 'i', 'p', 'z', 'j'])
+auto_detect_lang = detect_language(input_text)
+
+# Set detected language to the selectbox (if detected)
+if auto_detect_lang:
+    lang_map = {
+        'en': 'a',  # American English
+        'es': 'e',  # Spanish
+        'fr': 'f',  # French
+        'hi': 'h',  # Hindi
+        'it': 'i',  # Italian
+        'pt': 'p',  # Portuguese
+        'zh': 'z',  # Chinese
+        'ja': 'j'   # Japanese
+    }
+    lang_code = lang_map.get(auto_detect_lang, 'a')  # Default to English if not in map
+else:
+    lang_code = st.selectbox("Select Language", ['a', 'b', 'e', 'f', 'h', 'i', 'p', 'z', 'j'])
+
 voice = st.selectbox("Select Voice", ['af_alloy', 'af_aoede', 'af_bella', 'af_heart', 'af_jessica', 'af_kore', 'af_nicole', 'af_nova', 'af_river', 'af_sarah', 'af_sky', 
-                                     'am_adam', 'am_echo', 'am_eric', 'am_fenrir', 'am_liam', 'am_michael', 'am_onyx', 'am_puck', 'am_santa', 
-                                     'bf_alice', 'bf_emma', 'bf_isabella', 'bf_lily', 
-                                     'bm_daniel', 'bm_fable', 'bm_george', 'bm_lewis', 
-                                     'ef_dora', 
-                                     'em_alex', 'em_santa', 
-                                     'ff_siwis', 
-                                     'hf_alpha', 'hf_beta', 
-                                     'hm_omega', 'hm_psi', 
-                                     'if_sara', 
-                                     'im_nicola', 
-                                     'jf_alpha', 'jf_gongitsune', 'jf_nezumi', 'jf_tebukuro', 
-                                     'jm_kumo', 
-                                     'pf_dora', 
-                                     'pm_alex', 'pm_santa', 
-                                     'zf_xiaobei', 'zf_xiaoni', 'zf_xiaoxiao', 'zf_xiaoyi', 
-                                     'zm_yunjian', 'zm_yunxi', 'zm_yunxia', 'zm_yunyang'])
-speed = st.slider("Speech Speed", min_value=0.5, max_value=2.0, value=1.0, step=0.1)
-
-background_music = st.selectbox("Add Background Music", ['None', 'Calm', 'Energetic', 'Focus', 'Nature'])
+ 'am_adam', 'am_echo', 'am_eric', 'am_fenrir', 'am_liam', 'am_michael', 'am_onyx', 'am_puck', 'am_santa', 
+ 'bf_alice', 'bf_emma', 'bf_isabella', 'bf_lily', 
+ 'bm_daniel', 'bm_fable', 'bm_george', 'bm_lewis', 
+ 'ef_dora', 
+ 'em_alex', 'em_santa', 
+ 'ff_siwis', 
+ 'hf_alpha', 'hf_beta', 
+ 'hm_omega', 'hm_psi', 
+ 'if_sara', 
+ 'im_nicola', 
+ 'jf_alpha', 'jf_gongitsune', 'jf_nezumi', 'jf_tebukuro', 
+ 'jm_kumo', 
+ 'pf_dora', 
+ 'pm_alex', 'pm_santa', 
+ 'zf_xiaobei', 'zf_xiaoni', 'zf_xiaoxiao', 'zf_xiaoyi', 
+ 'zm_yunjian', 'zm_yunxi', 'zm_yunxia', 'zm_yunyang'])
+
+speed = st.slider("Speed", min_value=0.5, max_value=2.0, value=1.0, step=0.1)
 
 # Initialize the TTS pipeline with user-selected language
 pipeline = KPipeline(lang_code=lang_code)
 
-# Function to generate audio with background music
-def generate_audio_with_music(text, lang_code, voice, speed, music_type):
+# Generate Audio function
+def generate_audio(text, lang_code, voice, speed):
     generator = pipeline(text, voice=voice, speed=speed, split_pattern=r'\n+')
     for i, (gs, ps, audio) in enumerate(generator):
         audio_data = audio
-        
-        # Simulate adding background music (using silence as placeholder)
-        if music_type != 'None':
-            st.write(f"Adding {music_type} background music...")
-            # Add some background music (could be a placeholder for real music loading)
-            music = io.BytesIO()
-            music.write(b'\0' * 500000)  # Placeholder: Replace with actual music blending logic
-
         # Save audio to in-memory buffer
         buffer = io.BytesIO()
-        sf.write(buffer, audio_data, 24000, format='WAV')
+        # Explicitly specify format as WAV
+        sf.write(buffer, audio_data, 24000, format='WAV')  # Add 'format="WAV"'
         buffer.seek(0)
         return buffer
 
-# Generate Audio Button
+# Generate and display the audio file
 if st.button('Generate Audio'):
-    st.write("Generating speech with background music...")
-    audio_buffer = generate_audio_with_music(input_text, lang_code, voice, speed, background_music)
+    st.write("Generating speech...")
+    audio_buffer = generate_audio(input_text, lang_code, voice, speed)
     
-    # Display audio player in the app
+    # Display Audio player in the app
     st.audio(audio_buffer, format='audio/wav')
 
-    # Option to download the audio file
+    # Optional: Save the generated audio file for download
     st.download_button(
         label="Download Audio",
         data=audio_buffer,
-        file_name="generated_speech_with_music.wav",
+        file_name="generated_speech.wav",
         mime="audio/wav"
     )
 
-    # Speech Progress Feedback: Add voice feedback or something fun
-    st.write("Would you like to hear the audio again?")
-    if st.button('Replay Audio'):
+    # Interactive Voice Feedback
+    feedback = st.radio("Do you want to hear it again?", ('No', 'Yes'))
+    if feedback == 'Yes':
+        st.write("Replaying the generated speech...")
         st.audio(audio_buffer, format='audio/wav')