Update src/modules/voice/voice.service.ts

src/modules/voice/voice.service.ts

@@ -1,137 +1,3 @@
-// import { config } from "@/config";
-// import * as googleTTS from 'google-tts-api';
-// import axios from 'axios';
-// import FormData from 'form-data';
-
-// export class VoiceService {
-
-//     async transcribeAudio(audioBuffer: Buffer): Promise<any> {
-//         try {
-//             // Detect file type from magic bytes
-//             const ext = this.detectAudioType(audioBuffer);
-//             console.log(`[STT] Received ${audioBuffer.length} bytes, detected type: ${ext}`);
-
-//             const form = new FormData();
-//             form.append('file', audioBuffer, {
-//                 filename: `audio.${ext}`,
-//                 contentType: this.getMimeType(ext),
-//             });
-//             form.append('model', 'whisper-large-v3');
-//             // Remove hardcoded 'ur' to allow auto-detection for English/Urdu bilingual support
-//             // form.append('language', 'ur'); 
-//             form.append('temperature', '0');
-//             form.append('response_format', 'json');
-
-//             const response = await axios.post(
-//                 'https://api.groq.com/openai/v1/audio/transcriptions',
-//                 form,
-//                 {
-//                     headers: {
-//                         'Authorization': `Bearer ${config.llmModel.groqApiKey}`,
-//                         ...form.getHeaders(),
-//                     },
-//                     maxContentLength: Infinity,
-//                     maxBodyLength: Infinity,
-//                 }
-//             );
-
-//             const text = response.data?.text || '';
-
-//             if (!text.trim()) {
-//                 console.warn('[STT] Empty transcription returned');
-//                 return { error: "STT returned empty text" };
-//             }
-
-//             console.log(`\n\x1b[32m========== USER INPUT ==========\x1b[0m`);
-//             console.log(`\x1b[32m[User]: ${text}\x1b[0m`);
-//             console.log(`\x1b[32m================================\x1b[0m\n`);
-            
-//             return { text };
-
-//         } catch (error: any) {
-//             if (error.response) {
-//                 console.error(`[STT] Groq API Error ${error.response.status}:`, error.response.data);
-//                 return { error: "STT Failed", status: error.response.status, details: error.response.data };
-//             }
-//             console.error("Groq Transcribe Error:", error.message || error);
-//             return { error: "STT Failed", details: error.message || String(error) };
-//         }
-//     }
-
-//     private detectAudioType(buffer: Buffer): string {
-//         // WebM magic bytes: 1A 45 DF A3
-//         if (buffer.length >= 4 && buffer[0] === 0x1A && buffer[1] === 0x45 && buffer[2] === 0xDF && buffer[3] === 0xA3) {
-//             return 'webm';
-//         }
-//         // WAV magic bytes: RIFF....WAVE
-//         if (buffer.length >= 12 && buffer.toString('ascii', 0, 4) === 'RIFF' && buffer.toString('ascii', 8, 12) === 'WAVE') {
-//             return 'wav';
-//         }
-//         // MP4/M4A magic bytes: ....ftyp
-//         if (buffer.length >= 8 && buffer.toString('ascii', 4, 8) === 'ftyp') {
-//             return 'm4a';
-//         }
-//         // OGG magic bytes: OggS
-//         if (buffer.length >= 4 && buffer.toString('ascii', 0, 4) === 'OggS') {
-//             return 'ogg';
-//         }
-//         // Default to webm (most common from browser)
-//         return 'webm';
-//     }
-
-//     private getMimeType(ext: string): string {
-//         const mimeTypes: Record<string, string> = {
-//             'webm': 'audio/webm',
-//             'wav': 'audio/wav',
-//             'm4a': 'audio/mp4',
-//             'ogg': 'audio/ogg',
-//             'mp3': 'audio/mpeg',
-//         };
-//         return mimeTypes[ext] || 'application/octet-stream';
-//     }
-
-//     async generateTTSAudio(text: string): Promise<Buffer> {
-//         try {
-//             console.log(`[TTS] Requesting Google TTS for: "${text.substring(0, 30)}..." (Length: ${text.length})`);
-            
-//             // Detect if Urdu script present to set language
-//             const isUrdu = /[\u0600-\u06FF]/.test(text);
-//             const lang = isUrdu ? 'ur' : 'en';
-
-//             // Split text into chunks of 200 characters (Google TTS limit)
-//             const chunks = googleTTS.getAllAudioUrls(text, {
-//                 lang: lang,
-//                 slow: false,
-//                 host: 'https://translate.google.com',
-//                 splitPunct: '. ',
-//             });
-
-//             console.log(`[TTS] Fetching ${chunks.length} audio chunks from Google...`);
-            
-//             const bufferChunks: Buffer[] = [];
-//             // Run requests in parallel to reduce delay
-//             const requests = chunks.map(chunk => 
-//                 axios.get(chunk.url, { responseType: 'arraybuffer' })
-//             );
-            
-//             const responses = await Promise.all(requests);
-//             for (const response of responses) {
-//                 bufferChunks.push(Buffer.from(response.data));
-//             }
-
-//             const finalBuffer = Buffer.concat(bufferChunks);
-//             console.log(`[TTS] Successfully received and concatenated ${chunks.length} chunks (${finalBuffer.byteLength} bytes)`);
-//             return finalBuffer;
-
-//         } catch (error: any) {
-//             console.error("Google TTS Service Error:", error.message || error);
-//             throw error;
-//         }
-//     }
-// }
-
-
-
 import { config } from "@/config";
 import * as googleTTS from 'google-tts-api';
 import axios from 'axios';
@@ -150,7 +16,7 @@ export class VoiceService {
                 filename: `audio.${ext}`,
                 contentType: this.getMimeType(ext),
             });
-            form.append('model', 'whisper-large-v3-turbo');
+            form.append('model', 'whisper-large-v3');
             // Remove hardcoded 'ur' to allow auto-detection for English/Urdu bilingual support
             // form.append('language', 'ur'); 
             form.append('temperature', '0');
@@ -224,14 +90,13 @@ export class VoiceService {
         return mimeTypes[ext] || 'application/octet-stream';
     }
 
-    async generateTTSAudio(text: string, language?: string): Promise<Buffer> {
+    async generateTTSAudio(text: string): Promise<Buffer> {
         try {
             console.log(`[TTS] Requesting Google TTS for: "${text.substring(0, 30)}..." (Length: ${text.length})`);
             
-            // Use explicit language if provided, otherwise auto-detect from text
+            // Detect if Urdu script present to set language
             const isUrdu = /[\u0600-\u06FF]/.test(text);
-            const lang = language || (isUrdu ? 'ur' : 'en');
-            console.log(`[TTS] Language: ${lang} (explicit: ${!!language}, urduScript: ${isUrdu})`);
+            const lang = isUrdu ? 'ur' : 'en';
 
             // Split text into chunks of 200 characters (Google TTS limit)
             const chunks = googleTTS.getAllAudioUrls(text, {
@@ -264,3 +129,138 @@ export class VoiceService {
         }
     }
 }
+
+
+
+// import { config } from "@/config";
+// import * as googleTTS from 'google-tts-api';
+// import axios from 'axios';
+// import FormData from 'form-data';
+
+// export class VoiceService {
+
+//     async transcribeAudio(audioBuffer: Buffer): Promise<any> {
+//         try {
+//             // Detect file type from magic bytes
+//             const ext = this.detectAudioType(audioBuffer);
+//             console.log(`[STT] Received ${audioBuffer.length} bytes, detected type: ${ext}`);
+
+//             const form = new FormData();
+//             form.append('file', audioBuffer, {
+//                 filename: `audio.${ext}`,
+//                 contentType: this.getMimeType(ext),
+//             });
+//             form.append('model', 'whisper-large-v3-turbo');
+//             // Remove hardcoded 'ur' to allow auto-detection for English/Urdu bilingual support
+//             // form.append('language', 'ur'); 
+//             form.append('temperature', '0');
+//             form.append('response_format', 'json');
+
+//             const response = await axios.post(
+//                 'https://api.groq.com/openai/v1/audio/transcriptions',
+//                 form,
+//                 {
+//                     headers: {
+//                         'Authorization': `Bearer ${config.llmModel.groqApiKey}`,
+//                         ...form.getHeaders(),
+//                     },
+//                     maxContentLength: Infinity,
+//                     maxBodyLength: Infinity,
+//                 }
+//             );
+
+//             const text = response.data?.text || '';
+
+//             if (!text.trim()) {
+//                 console.warn('[STT] Empty transcription returned');
+//                 return { error: "STT returned empty text" };
+//             }
+
+//             console.log(`\n\x1b[32m========== USER INPUT ==========\x1b[0m`);
+//             console.log(`\x1b[32m[User]: ${text}\x1b[0m`);
+//             console.log(`\x1b[32m================================\x1b[0m\n`);
+            
+//             return { text };
+
+//         } catch (error: any) {
+//             if (error.response) {
+//                 console.error(`[STT] Groq API Error ${error.response.status}:`, error.response.data);
+//                 return { error: "STT Failed", status: error.response.status, details: error.response.data };
+//             }
+//             console.error("Groq Transcribe Error:", error.message || error);
+//             return { error: "STT Failed", details: error.message || String(error) };
+//         }
+//     }
+
+//     private detectAudioType(buffer: Buffer): string {
+//         // WebM magic bytes: 1A 45 DF A3
+//         if (buffer.length >= 4 && buffer[0] === 0x1A && buffer[1] === 0x45 && buffer[2] === 0xDF && buffer[3] === 0xA3) {
+//             return 'webm';
+//         }
+//         // WAV magic bytes: RIFF....WAVE
+//         if (buffer.length >= 12 && buffer.toString('ascii', 0, 4) === 'RIFF' && buffer.toString('ascii', 8, 12) === 'WAVE') {
+//             return 'wav';
+//         }
+//         // MP4/M4A magic bytes: ....ftyp
+//         if (buffer.length >= 8 && buffer.toString('ascii', 4, 8) === 'ftyp') {
+//             return 'm4a';
+//         }
+//         // OGG magic bytes: OggS
+//         if (buffer.length >= 4 && buffer.toString('ascii', 0, 4) === 'OggS') {
+//             return 'ogg';
+//         }
+//         // Default to webm (most common from browser)
+//         return 'webm';
+//     }
+
+//     private getMimeType(ext: string): string {
+//         const mimeTypes: Record<string, string> = {
+//             'webm': 'audio/webm',
+//             'wav': 'audio/wav',
+//             'm4a': 'audio/mp4',
+//             'ogg': 'audio/ogg',
+//             'mp3': 'audio/mpeg',
+//         };
+//         return mimeTypes[ext] || 'application/octet-stream';
+//     }
+
+//     async generateTTSAudio(text: string, language?: string): Promise<Buffer> {
+//         try {
+//             console.log(`[TTS] Requesting Google TTS for: "${text.substring(0, 30)}..." (Length: ${text.length})`);
+            
+//             // Use explicit language if provided, otherwise auto-detect from text
+//             const isUrdu = /[\u0600-\u06FF]/.test(text);
+//             const lang = language || (isUrdu ? 'ur' : 'en');
+//             console.log(`[TTS] Language: ${lang} (explicit: ${!!language}, urduScript: ${isUrdu})`);
+
+//             // Split text into chunks of 200 characters (Google TTS limit)
+//             const chunks = googleTTS.getAllAudioUrls(text, {
+//                 lang: lang,
+//                 slow: false,
+//                 host: 'https://translate.google.com',
+//                 splitPunct: '. ',
+//             });
+
+//             console.log(`[TTS] Fetching ${chunks.length} audio chunks from Google...`);
+            
+//             const bufferChunks: Buffer[] = [];
+//             // Run requests in parallel to reduce delay
+//             const requests = chunks.map(chunk => 
+//                 axios.get(chunk.url, { responseType: 'arraybuffer' })
+//             );
+            
+//             const responses = await Promise.all(requests);
+//             for (const response of responses) {
+//                 bufferChunks.push(Buffer.from(response.data));
+//             }
+
+//             const finalBuffer = Buffer.concat(bufferChunks);
+//             console.log(`[TTS] Successfully received and concatenated ${chunks.length} chunks (${finalBuffer.byteLength} bytes)`);
+//             return finalBuffer;
+
+//         } catch (error: any) {
+//             console.error("Google TTS Service Error:", error.message || error);
+//             throw error;
+//         }
+//     }
+// }