Spaces:
Runtime error
Runtime error
Commit History
-
Add 899999999999999999999999999999
5065491
-
Add 8999999999999999999999999999
ce2d17d -
Add 89999999999999999999999999
2b377e5 -
Add 89999999999999999999999999
4a1bc78 -
Add 8999999999999999999999999
9341027
-
Enhance voice bot with comprehensive multi-language and intelligent query handling
dad5387
-
Add 89999999999999999999999
82583bd -
Add 8999999999999999999999
15b123d -
Add 899999999999999999999
f9f45e2 -
Add 89999999999999999999
f139d4e
-
Add 8999999999999999999
c3ade54 -
Add 89999999999999999
36ccb7e -
Add 899999999999999
061a93c -
Add 89999999999999
a0b0f78 -
Add 899999999999
899843c
-
Add 89999999999
ed60e36 -
Add 8999999999
d15e519
-
Add 8999999
b297e4e
-
Send TTS audio message for frontend playback
bef54cd -
Send TTS audio as binary WebSocket message for frontend playback
5060335 -
🔊 Enable voice responses by default for Voice Bot
- Change default response_mode from 'text' to 'both' for voice+text responses - Enable voice_enabled by default since this is primarily a voice bot - This ensures users get both text and audio responses from the bot - Fixes issue where bot was generating responses but not delivering voice output1eb2c60
-
🔧 Fix WebSocket response message type mismatch
- Change message type from 'message_response' to 'llm_response' to match frontend expectations - Change message field from 'message' to 'text' for consistency - Fix scenario analysis response format - This resolves the issue where bot responses weren't displaying to users943f386 -
🔧 Improve FFmpeg installation and browser ASR handling
- Update Dockerfile to explicitly install FFmpeg system dependency - Enhance packages.txt with additional build dependencies - Add intelligent browser-native ASR support in WebSocket handler - Handle transcription from browser when server ASR is unavailable - Improve error messaging for different ASR modes75546b0 -
🎤 Enhance voice bot with multi-language support and fix MediaRecorder issues
- Fix WebSocket message handling to properly process voice messages with type field - Add comprehensive language support (English, Hindi, Hinglish) for ASR and TTS - Enhance error handling for voice processing - Fix KeyError: 'text' by improving message validation - Add language parameter support in voice_service.py speech_to_text method - Improve audio data processing with base64 encoding support8a78e3e -
Add comprehensive WebSocket error handling and message validation
- Add proper error handling for missing 'type' field in messages - Validate message format and log detailed debugging information - Handle connection and knowledge base requests properly - Add graceful error responses for invalid messages - This should fix the 'type' and 'text' KeyError messages in logs102aa44 -
Fix WebSocket connection hanging - send connection_successful message
- Change greeting message type from 'message_response' to 'connection_successful' - Frontend expects connection_successful to complete WebSocket handshake - This should fix the 'loading' state that never completes - Users will now see the welcome message and can start chatting08a93b9 -
Fix RAG search issues and add document database
- Replace LangChain tool call with direct search_documents function - Fix 'parent_run_id' error by using proper function interface - Add complete lancedb_data with Rajasthan documents and other knowledge bases - This should resolve document search failures and improve response speed1921353 -
Fix slow response and improve RAG search performance
- Fix search_government_docs call from .ainvoke() to direct function call - Add detailed logging to track search and LLM response timing - Add error handling for document search failures - This should fix the slow loading and hanging issuesbbe8a73 -
Fix NoneType error in provider selection
- Fix choose_llm_provider call to pass message instead of complexity - Add null check for provider value to prevent attribute errors - Provider selection now works correctly in WebSocket handlera174809
-
Deploy clean Voice Bot backend to HF Spaces
🚀 Features: - FastAPI application optimized for HF Spaces (port 7860) - Voice processing with ASR and TTS - LangChain-powered RAG system for document search - WebSocket support for real-time communication - JWT authentication - Hybrid LLM service (Gemini + Groq) - Docker configuration with health checks - Clean project structure without deployment artifacts ✅ Ready for HF Spaces deploymentcf02b2b