voice

mnemo_cards_admin/README.md

@@ -2,6 +2,20 @@
 
 This template provides a minimal setup to get React working in Vite with HMR and some ESLint rules.
 
+## Environment Variables
+
+The following environment variables can be configured:
+
+- `VITE_API_BASE_URL` - Base URL for the API (default: `https://api.mnemo-cards.online`)
+- `VITE_VOICE_SERVICE_URL` - Base URL for the voice-service TTS API (default: `https://voice-service.mnemo-cards.online`)
+
+Create a `.env` file in the root directory to override these values:
+
+```env
+VITE_API_BASE_URL=https://api.mnemo-cards.online
+VITE_VOICE_SERVICE_URL=https://voice-service.mnemo-cards.online
+```
+
 Currently, two official plugins are available:
 
 - [@vitejs/plugin-react](https://github.com/vitejs/vite-plugin-react/blob/main/packages/plugin-react) uses [Babel](https://babeljs.io/) (or [oxc](https://oxc.rs) when used in [rolldown-vite](https://vite.dev/guide/rolldown)) for Fast Refresh

mnemo_cards_admin/src/api/media.ts

@@ -58,11 +58,15 @@ export const mediaApi = {
   /**
    * Upload voice audio file to MinIO
    * @param file Audio file to upload
+   * @param objectId Optional object ID (hash) for caching. If provided, will use this as the object ID in MinIO
    * @returns Object ID and presigned URL
    */
-  async uploadVoice(file: File): Promise<UploadResponse> {
+  async uploadVoice(file: File, objectId?: string): Promise<UploadResponse> {
     const formData = new FormData()
     formData.append('file', file)
+    if (objectId) {
+      formData.append('objectId', objectId)
+    }
 
     const response = await adminApiClient.post<UploadResponse>(
       '/api/v2/media/upload/voice',

mnemo_cards_admin/src/api/voiceService.ts

@@ -0,0 +1,30 @@
+import { adminApiClient } from './client'
+
+export interface GenerateTTSResponse {
+  objectId: string
+  url: string // Presigned URL for preview
+}
+
+export const voiceServiceApi = {
+  /**
+   * Generate TTS audio from text via backend API
+   * Backend calls voice-service, which handles caching, and saves to MinIO
+   * @param text Text to convert to speech
+   * @param voice Voice identifier (optional, defaults to 'alloy')
+   * @returns Object ID and presigned URL
+   */
+  async generateTTS(
+    text: string,
+    voice?: string
+  ): Promise<GenerateTTSResponse> {
+    const response = await adminApiClient.post<GenerateTTSResponse>(
+      '/api/v2/media/generate-tts',
+      {
+        text,
+        voice: voice || 'alloy',
+      }
+    )
+
+    return response.data
+  },
+}

mnemo_cards_admin/src/components/CardEditorPreview.tsx

@@ -349,7 +349,11 @@ export function CardEditorPreview({
         {/* Voice Controls - only show if cardId exists */}
         {cardId && (
           <div className="space-y-2">
-            <CardVoicesManager cardId={cardId} disabled={disabled} />
+            <CardVoicesManager 
+              cardId={cardId} 
+              cardOriginal={formData.original}
+              disabled={disabled} 
+            />
           </div>
         )}
       </div>

mnemo_cards_admin/src/components/CardVoicesManager.tsx

@@ -6,19 +6,26 @@ import { AudioUpload } from '@/components/ui/audio-upload'
 import { Button } from '@/components/ui/button'
 import { Label } from '@/components/ui/label'
 import { Badge } from '@/components/ui/badge'
-import { X, Plus, Music, Play, Pause } from 'lucide-react'
+import { X, Plus, Music, Play, Pause, Volume2, Loader2 } from 'lucide-react'
 import type { AxiosError } from 'axios'
+import { voiceServiceApi } from '@/api/voiceService'
 
 interface CardVoicesManagerProps {
   cardId: string
+  cardOriginal?: string // Text from card's original field for TTS generation
   disabled?: boolean
 }
 
-export function CardVoicesManager({ cardId, disabled = false }: CardVoicesManagerProps) {
+export function CardVoicesManager({ 
+  cardId, 
+  cardOriginal, 
+  disabled = false 
+}: CardVoicesManagerProps) {
   const queryClient = useQueryClient()
   const [showAddForm, setShowAddForm] = useState(false)
   const [newAudio, setNewAudio] = useState<string | undefined>(undefined)
   const [newLanguage, setNewLanguage] = useState('en')
+  const [isGeneratingTTS, setIsGeneratingTTS] = useState(false)
 
   // Load voices for the card
   const { data: voicesData, isLoading, refetch } = useQuery({
@@ -88,6 +95,50 @@ export function CardVoicesManager({ cardId, disabled = false }: CardVoicesManage
     }
   }
 
+  // Generate TTS audio for card
+  const handleGenerateTTS = async () => {
+    if (!cardOriginal || !cardOriginal.trim()) {
+      toast.error('Card text is required for TTS generation')
+      return
+    }
+
+    if (!cardId) {
+      toast.error('Card ID is required')
+      return
+    }
+
+    setIsGeneratingTTS(true)
+    try {
+      const text = cardOriginal.trim()
+      // Use default voice 'alloy' for TTS generation
+      // Language (newLanguage) is used only when saving the voice to the card
+      const voice = 'alloy'
+      
+      console.log('🔄 CardVoicesManager: Requesting TTS generation', { text, voice })
+      
+      // Generate TTS via backend API
+      // Backend calls voice-service (which handles caching) and saves to MinIO
+      const ttsResponse = await voiceServiceApi.generateTTS(text, voice)
+      console.log('✅ CardVoicesManager: TTS generated and saved to MinIO', { 
+        objectId: ttsResponse.objectId 
+      })
+
+      // Add voice to card using the objectId from backend
+      await addVoiceMutation.mutateAsync({
+        voiceUrl: ttsResponse.objectId,
+        language: newLanguage || 'en',
+      })
+    } catch (error) {
+      console.error('❌ CardVoicesManager: TTS generation failed', error)
+      const errorMessage = error instanceof Error 
+        ? error.message 
+        : 'Failed to generate TTS audio'
+      toast.error(`TTS generation failed: ${errorMessage}`)
+    } finally {
+      setIsGeneratingTTS(false)
+    }
+  }
+
   const voices = voicesData?.items || []
 
   return (
@@ -95,16 +146,39 @@ export function CardVoicesManager({ cardId, disabled = false }: CardVoicesManage
       <div className="flex items-center justify-between">
         <Label>Card Voices ({voices.length})</Label>
         {!showAddForm && (
-          <Button
+          <div className="flex items-center space-x-2">
-            type="button"
+            {cardOriginal && cardOriginal.trim() && (
-            variant="outline"
+              <Button
-            size="sm"
+                type="button"
-            onClick={() => setShowAddForm(true)}
+                variant="outline"
-            disabled={disabled}
+                size="sm"
-          >
+                onClick={handleGenerateTTS}
-            <Plus className="h-4 w-4 mr-2" />
+                disabled={disabled || isGeneratingTTS}
-            Add Voice
+              >
-          </Button>
+                {isGeneratingTTS ? (
+                  <>
+                    <Loader2 className="h-4 w-4 mr-2 animate-spin" />
+                    Generating...
+                  </>
+                ) : (
+                  <>
+                    <Volume2 className="h-4 w-4 mr-2" />
+                    Generate TTS
+                  </>
+                )}
+              </Button>
+            )}
+            <Button
+              type="button"
+              variant="outline"
+              size="sm"
+              onClick={() => setShowAddForm(true)}
+              disabled={disabled}
+            >
+              <Plus className="h-4 w-4 mr-2" />
+              Add Voice
+            </Button>
+          </div>
         )}
       </div>
 

mnemo_cards_backend/lib/api/v2/media_api_v2.dart

@@ -1,6 +1,7 @@
 import 'dart:convert';
 import 'dart:typed_data';
 
+import 'package:dio/dio.dart' as dio;
 import 'package:injectable/injectable.dart';
 import 'package:mnemo_cards_backend/api/authorize/acl_types.dart';
 import 'package:mnemo_cards_backend/api/authorize/access_service.dart';
@@ -18,6 +19,12 @@ part 'media_api_v2.g.dart';
 @lazySingleton
 class MediaApiV2 {
   final MinioService _minioService;
+  final dio.Dio _dio = dio.Dio();
+  // Voice service URL from environment or default
+  final String _voiceServiceUrl = const String.fromEnvironment(
+    'VOICE_SERVICE_URL',
+    defaultValue: 'https://voice-service.mnemo-cards.online',
+  );
 
   MediaApiV2(this._minioService);
 
@@ -117,6 +124,83 @@ class MediaApiV2 {
     }
   }
 
+  /// Parses multipart/form-data request and extracts file and optional objectId
+  Future<({
+    Uint8List bytes,
+    String contentType,
+    String? filename,
+    String? objectId,
+  })?> _parseMultipartFileWithObjectId(Request request) async {
+    try {
+      final multipart = request.multipart();
+      if (multipart == null) {
+        return null;
+      }
+
+      Uint8List? fileBytes;
+      String? fileContentType;
+      String? filename;
+      String? objectId;
+
+      await for (final part in multipart.parts) {
+        final contentDisposition = part.headers['content-disposition'] ?? '';
+        
+        if (contentDisposition.contains('name="file"')) {
+          // Read file bytes
+          final chunks = <List<int>>[];
+          await for (final chunk in part) {
+            chunks.add(chunk);
+          }
+          fileBytes = Uint8List.fromList(
+            chunks.expand((chunk) => chunk).toList(),
+          );
+
+          // Get content type from part headers
+          fileContentType =
+              part.headers['content-type'] ?? 'application/octet-stream';
+
+          // Extract filename from content-disposition header
+          final filenameMatch = RegExp(
+            r'filename="?([^"]+)"?',
+          ).firstMatch(contentDisposition);
+          filename = filenameMatch?.group(1);
+        } else if (contentDisposition.contains('name="objectId"')) {
+          // Read objectId field
+          final chunks = <List<int>>[];
+          await for (final chunk in part) {
+            chunks.add(chunk);
+          }
+          final objectIdBytes = Uint8List.fromList(
+            chunks.expand((chunk) => chunk).toList(),
+          );
+          objectId = utf8.decode(objectIdBytes).trim();
+          if (objectId.isEmpty) {
+            objectId = null;
+          }
+        } else {
+          // Read and discard other parts to ensure full request is consumed
+          await for (final _ in part) {
+            // Discard chunks from other parts
+          }
+        }
+      }
+
+      if (fileBytes == null) {
+        return null;
+      }
+
+      return (
+        bytes: fileBytes,
+        contentType: fileContentType!,
+        filename: filename,
+        objectId: objectId,
+      );
+    } catch (e) {
+      print('Error parsing multipart with objectId: $e');
+      return null;
+    }
+  }
+
   /// POST /api/v2/media/upload/card-image
   /// Upload card image to MinIO
   @Route.post('/media/upload/card-image')
@@ -242,6 +326,7 @@ class MediaApiV2 {
 
   /// POST /api/v2/media/upload/voice
   /// Upload voice audio file to MinIO
+  /// Supports optional objectId parameter for caching (hash-based)
   @Route.post('/media/upload/voice')
   @OpenApiRouteHttp()
   Future<Response> uploadVoice(Request request) async {
@@ -252,8 +337,8 @@ class MediaApiV2 {
         return authResponse;
       }
 
-      // Parse multipart file
+      // Parse multipart file with optional objectId
-      final fileData = await _parseMultipartFile(request);
+      final fileData = await _parseMultipartFileWithObjectId(request);
       if (fileData == null) {
         return _badRequest('No file provided or invalid multipart format');
       }
@@ -278,20 +363,41 @@ class MediaApiV2 {
         return _badRequest('File size exceeds maximum allowed size of 20MB');
       }
 
-      // Upload to MinIO
+      // Validate objectId if provided (should be UUID or MD5 hash - 32 hex chars)
-      final objectId = await _minioService.uploadFile(
+      String? objectId = fileData.objectId;
+      if (objectId != null) {
+        // Validate: UUID format (8-4-4-4-12) or MD5 hash (32 hex chars)
+        final isValidUuid = RegExp(
+          r'^[0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12}$',
+          caseSensitive: false,
+        ).hasMatch(objectId);
+        final isValidHash = RegExp(
+          r'^[0-9a-f]{32}$',
+          caseSensitive: false,
+        ).hasMatch(objectId);
+        
+        if (!isValidUuid && !isValidHash) {
+          return _badRequest(
+            'Invalid objectId format. Must be UUID or MD5 hash (32 hex characters)',
+          );
+        }
+      }
+
+      // Upload to MinIO with optional objectId
+      final finalObjectId = await _minioService.uploadFile(
         bucket: MinioConfig.voiceAudioBucket,
         bytes: fileData.bytes,
         contentType: fileData.contentType,
+        objectId: objectId,
       );
 
       // Generate presigned URL for preview
       final presignedUrl = await _minioService.getPresignedUrl(
         bucket: MinioConfig.voiceAudioBucket,
-        objectId: objectId,
+        objectId: finalObjectId,
       );
 
-      return _ok({'objectId': objectId, 'url': presignedUrl});
+      return _ok({'objectId': finalObjectId, 'url': presignedUrl});
     } catch (e, s) {
       print('Error uploading voice: $e\n$s');
       return _internalServerError('Failed to upload audio');
@@ -386,4 +492,78 @@ class MediaApiV2 {
       return _internalServerError('Failed to delete file');
     }
   }
+
+  /// POST /api/v2/media/generate-tts
+  /// Generate TTS audio via voice-service and save to MinIO
+  /// Voice-service handles its own caching, we save the result to our MinIO
+  @Route.post('/media/generate-tts')
+  @OpenApiRouteHttp()
+  Future<Response> generateTTS(Request request) async {
+    try {
+      // Check admin rights
+      final authResponse = await _ensureAdmin(request);
+      if (authResponse.statusCode != 200) {
+        return authResponse;
+      }
+
+      // Parse request body
+      final body = await request.readAsString();
+      final json = jsonDecode(body) as Map<String, dynamic>;
+      final text = json['text'] as String?;
+      final voice = json['voice'] as String?;
+
+      if (text == null || text.trim().isEmpty) {
+        return _badRequest('Text is required');
+      }
+
+      print(
+        '🔄 MediaApiV2: Requesting TTS from voice-service: text="${text.substring(0, text.length > 50 ? 50 : text.length)}", voice="${voice ?? 'default'}"',
+      );
+
+      // Call voice-service
+      final response = await _dio.post<Uint8List>(
+        '$_voiceServiceUrl/api/voice/tts',
+        data: {
+          'text': text,
+          if (voice != null) 'voice': voice,
+          'format': 'mp3',
+        },
+        options: dio.Options(
+          responseType: dio.ResponseType.bytes,
+          headers: {'Content-Type': 'application/json'},
+        ),
+      );
+
+      if (response.statusCode != 200 || response.data == null) {
+        return _internalServerError(
+          'Voice-service returned error: ${response.statusCode}',
+        );
+      }
+
+      final audioBytes = response.data!;
+      print(
+        '✅ MediaApiV2: Received TTS audio from voice-service: size=${audioBytes.length}',
+      );
+
+      // Save to MinIO (voice-service already cached it in its own storage)
+      final objectId = await _minioService.uploadFile(
+        bucket: MinioConfig.voiceAudioBucket,
+        bytes: audioBytes,
+        contentType: 'audio/mpeg',
+      );
+
+      // Generate presigned URL for preview
+      final presignedUrl = await _minioService.getPresignedUrl(
+        bucket: MinioConfig.voiceAudioBucket,
+        objectId: objectId,
+      );
+
+      print('✅ MediaApiV2: Saved TTS audio to MinIO: objectId=$objectId');
+
+      return _ok({'objectId': objectId, 'url': presignedUrl});
+    } catch (e, s) {
+      print('❌ MediaApiV2: Error generating TTS: $e\n$s');
+      return _internalServerError('Failed to generate TTS audio');
+    }
+  }
 }