Update app.py

app.py

@@ -6,6 +6,8 @@ from pymongo import MongoClient
 from bson import ObjectId
 from huggingface_hub import HfApi, HfFolder, delete_file
 import tempfile
+import base64
+import requests  # New import
 
 # Load environment variables
 load_dotenv()
@@ -57,12 +59,130 @@ def add_voice(category, name, is_free):
     # Check for duplicate voice name within the same category
     if voices_collection.find_one({"category": category, "voices.name": name}):
         st.error(f"Voice '{name}' already exists in category '{category}'.")
-        return
+        return False
     voices_collection.update_one(
         {"category": category},
         {"$push": {"voices": {"name": name, "free": is_free}}},
         upsert=True  # Ensure the category exists
     )
+    return True
+
+# Function to add a new voice with base64
+def add_voice_base64(category, name, is_free, audio_file):
+    # Check for duplicate voice name within the same category
+    if voices_collection.find_one({"category": category, "voices.name": name}):
+        st.error(f"Voice '{name}' already exists in category '{category}'.")
+        return False
+    
+    # Encode the audio file to base64
+    audio_base64 = base64.b64encode(audio_file.getvalue()).decode('utf-8')
+    
+    # Update the database with the new voice
+    voices_collection.update_one(
+        {"category": category},
+        {"$push": {"voices": {"name": name, "free": is_free, "base64": audio_base64}}},
+        upsert=True  # Ensure the category exists
+    )
+    
+    return True
+
+# Function to remove a voice
+def remove_voice(category, name):
+    # Find the voice to determine storage type
+    voice_doc = voices_collection.find_one({"category": category, "voices.name": name})
+    if not voice_doc:
+        st.error(f"Voice '{name}' not found in category '{category}'.")
+        return False
+    voice_data = next((v for v in voice_doc['voices'] if v['name'] == name), None)
+    if not voice_data:
+        st.error(f"Voice '{name}' not found in category '{category}'.")
+        return False
+    
+    # Remove from MongoDB
+    result = voices_collection.update_one(
+        {"category": category},
+        {"$pull": {"voices": {"name": name}}}
+    )
+    
+    # If removal was successful, handle deletion from Hugging Face if applicable
+    if result.modified_count > 0:
+        if 'base64' not in voice_data:
+            # Voice is stored on Hugging Face
+            try:
+                api = HfApi()
+                delete_file(
+                    path_in_repo=f"voices/{name}.mp3",
+                    repo_id="nikkmitra/clone",
+                    repo_type="space",
+                    token=hf_token
+                )
+                st.success(f"Deleted {name}'s voice file from Hugging Face repository")
+            except Exception as e:
+                st.warning(f"Failed to delete {name}'s voice file from Hugging Face: {str(e)}")
+        else:
+            # Voice was stored as base64 in MongoDB
+            st.success(f"Removed {name} from the '{category}' category in the database")
+    return result.modified_count > 0
+
+# Function to update all voices by adding base64 field
+def update_all_voices():
+    st.info("Starting the bulk update process. This may take a while depending on the number of voices.")
+    
+    # Fetch all categories and their voices
+    all_categories = voices_collection.find()
+    
+    total_voices = 0
+    updated_voices = 0
+    failed_voices = []
+    
+    for category in all_categories:
+        category_name = category['category']
+        voices = category.get('voices', [])
+        
+        for voice in voices:
+            total_voices += 1
+            voice_name = voice['name']
+            
+            # Check if 'base64' field already exists
+            if 'base64' in voice:
+                continue  # Skip voices already updated
+            
+            # Assume the voice is stored on Hugging Face
+            file_path = f"voices/{voice_name}.mp3"
+            repo_id = "nikkmitra/clone"  # Update this if different
+            url = f"https://huggingface.co/{repo_id}/resolve/main/{file_path}"
+            
+            try:
+                # Download the audio file from Hugging Face
+                response = requests.get(url, headers={"Authorization": f"Bearer {hf_token}"})
+                if response.status_code == 200:
+                    audio_data = response.content
+                    
+                    # Encode to base64
+                    audio_base64 = base64.b64encode(audio_data).decode('utf-8')
+                    
+                    # Update the MongoDB document
+                    voices_collection.update_one(
+                        {"category": category_name, "voices.name": voice_name},
+                        {"$set": {"voices.$.base64": audio_base64}}
+                    )
+                    
+                    updated_voices += 1
+                else:
+                    failed_voices.append((category_name, voice_name, f"HTTP {response.status_code}"))
+            except Exception as e:
+                failed_voices.append((category_name, voice_name, str(e)))
+    
+    # Summary of the update process
+    st.success(f"Bulk update completed. Total voices processed: {total_voices}")
+    st.success(f"Voices successfully updated with base64: {updated_voices}")
+    
+    if failed_voices:
+        st.error(f"Failed to update {len(failed_voices)} voices:")
+        for fail in failed_voices:
+            st.error(f"Category: {fail[0]}, Voice: {fail[1]}, Reason: {fail[2]}")
+    else:
+        st.success("All voices updated successfully!")
 
 # Streamlit app
 st.title("Mitra Voices")
@@ -76,31 +196,57 @@ st.header("Upload Voice")
 
 if categories:
     # Select category from dropdown
-    selected_category = st.selectbox("Select Category", categories, help="Choose the category to which you want to add the voice")
+    selected_category = st.selectbox(
+        "Select Category",
+        categories,
+        help="Choose the category to which you want to add the voice"
+    )
     
     # Input for voice name
-    voice_name = st.text_input("Voice Name", placeholder="Enter the name of the voice")
+    voice_name = st.text_input(
+        "Voice Name",
+        placeholder="Enter the name of the voice"
+    )
     
     # File uploader for MP3 file
     st.write("Please upload an MP3 file of the voice:")
-    audio_file = st.file_uploader("Choose an MP3 file", type="mp3", help="Select an MP3 file of the voice")
+    audio_file = st.file_uploader(
+        "Choose an MP3 file",
+        type="mp3",
+        help="Select an MP3 file of the voice"
+    )
+    
+    # Storage method selection
+    storage_method = st.radio(
+        "Select Storage Method",
+        ("Upload to Hugging Face", "Store as Base64 in MongoDB"),
+        help="Choose how you want to store the voice file."
+    )
     
     if selected_category and voice_name and audio_file:
         st.audio(audio_file, format='audio/mp3')
         st.write("Preview the uploaded audio:")
 
-        if st.button("Upload Voice", type="primary"):
-            if upload_to_huggingface(audio_file, voice_name):
-                st.success(f"✅ Uploaded {voice_name}'s voice to Hugging Face repository")
-                
-                # Add the new voice to the selected category in the database
-                add_voice(selected_category, voice_name, False)  # Assuming new uploads are not free by default
-                st.success(f"✅ Added {voice_name} to the '{selected_category}' category in the database")
-                
-                # Clear the form by rerunning the app
-                st.rerun()
-            else:
-                st.error("❌ Failed to upload the voice file")
+        if storage_method == "Upload to Hugging Face":
+            if st.button("Upload Voice", type="primary"):
+                if upload_to_huggingface(audio_file, voice_name):
+                    st.success(f"✅ Uploaded {voice_name}'s voice to Hugging Face repository")
+                    
+                    # Add the new voice to the selected category in the database
+                    if add_voice(selected_category, voice_name, False):  # Assuming new uploads are not free by default
+                        st.success(f"✅ Added {voice_name} to the '{selected_category}' category in the database")
+                    
+                    # Clear the form by rerunning the app
+                    st.rerun()
+                else:
+                    st.error("❌ Failed to upload the voice file")
+        elif storage_method == "Store as Base64 in MongoDB":
+            if st.button("Store Voice", type="primary"):
+                if add_voice_base64(selected_category, voice_name, False, audio_file):
+                    st.success(f"✅ Stored {voice_name}'s voice in the '{selected_category}' category in the database as base64")
+                    st.rerun()
+                else:
+                    st.error("❌ Failed to store the voice file as base64")
     elif selected_category and (voice_name or audio_file):
         if not voice_name:
             st.info("Please enter the voice name to proceed")
@@ -112,43 +258,25 @@ else:
 # Add a separator
 st.markdown("---")
 
-# Function to clear the database
-def clear_database():
-    voices_collection.delete_many({})
-
 # Button to clear the database
 if st.button("Clear Database"):
     clear_database()
     st.success("Database cleared successfully.")
     st.rerun()
 
-# Function to remove a voice
-def remove_voice(category, name):
-    # Remove from MongoDB
-    result = voices_collection.update_one(
-        {"category": category},
-        {"$pull": {"voices": {"name": name}}}
-    )
-    
-    # If the voice was successfully removed from MongoDB, also delete from Hugging Face
-    if result.modified_count > 0:
-        try:
-            api = HfApi()
-            delete_file(
-                path_in_repo=f"voices/{name}.mp3",
-                repo_id="nikkmitra/clone",
-                repo_type="space",
-                token=hf_token
-            )
-            st.success(f"Deleted {name}'s voice file from Hugging Face repository")
-        except Exception as e:
-            st.warning(f"Failed to delete {name}'s voice file from Hugging Face: {str(e)}")
-    
-    return result.modified_count > 0
+# Bulk Update Section
+st.markdown("---")
+st.header("Bulk Update Voices")
+
+if st.button("Update All Voices"):
+    update_all_voices()
 
 # Add new voice category
 st.header("Add New Voice Category")
-new_category_name = st.text_input("New Category Name", placeholder="Enter new category name")
+new_category_name = st.text_input(
+    "New Category Name",
+    placeholder="Enter new category name"
+)
 
 if st.button("Add Category"):
     if new_category_name:
@@ -172,6 +300,15 @@ for category in all_voices:
         col1, col2, col3 = st.columns([3, 1, 1])
         with col1:
             st.write(f"{voice['name']} ({'Free' if voice['free'] else 'Premium'})")
+            if 'base64' in voice:
+                try:
+                    audio_data = base64.b64decode(voice['base64'])
+                    st.audio(audio_data, format='audio/mp3')
+                except Exception as e:
+                    st.error(f"Failed to load audio for {voice['name']}: {str(e)}")
+            else:
+                # If stored on Hugging Face, provide a link or fetch directly
+                st.write("🔗 Stored on Hugging Face")
         with col2:
             if st.button("Remove", key=f"remove_{category['category']}_{voice['name']}_{i}"):
                 if remove_voice(category['category'], voice['name']):
@@ -186,5 +323,3 @@ for category in all_voices:
                     {"$set": {"voices.$.free": not voice['free']}}
                 )
                 st.rerun()
-
-# Note: We don't need to explicitly close the MongoDB connection when using Streamlit