File size: 9,581 Bytes
3370983 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 |
"""
Voice Screener API Router.
Handles voice screening sessions, configuration, and audio/transcript saving.
"""
import logging
import os
from typing import Optional
from pathlib import Path
from fastapi import APIRouter, HTTPException, Query
from pydantic import BaseModel
from src.backend.agents.voice_screening.session_service import (
get_session_config,
save_voice_screening_session
)
from src.backend.agents.voice_screening.audio_processor import combine_and_export_audio
logger = logging.getLogger(__name__)
router = APIRouter()
# Request/Response Models
class CreateSessionRequest(BaseModel):
candidate_id: str
class CreateSessionResponse(BaseModel):
session_id: str
candidate_name: str
job_title: str
message: str
class SessionConfigResponse(BaseModel):
candidate_name: str
job_title: str
instructions: str
questions: list[str]
config: dict
class SaveSessionRequest(BaseModel):
session_id: str
candidate_id: str
transcript_text: str
proxy_token: str # Token to retrieve audio chunks from proxy
class SaveSessionResponse(BaseModel):
audio_file_path: Optional[str]
message: str
@router.post("/session/create", response_model=CreateSessionResponse)
async def create_session(request: CreateSessionRequest):
"""
Create a new voice screening session for a candidate.
Args:
request: Contains candidate_id
Returns:
Session information including session_id
"""
try:
import uuid
# Generate session ID
session_id = str(uuid.uuid4())
# Get session config (validates candidate exists)
config = get_session_config(request.candidate_id)
logger.info(f"Created session {session_id} for candidate {request.candidate_id}")
return CreateSessionResponse(
session_id=session_id,
candidate_name=config["candidate_name"],
job_title=config["job_title"],
message="Session created successfully"
)
except ValueError as e:
raise HTTPException(status_code=404, detail=str(e))
except Exception as e:
logger.error(f"Error creating session: {e}", exc_info=True)
raise HTTPException(status_code=500, detail=f"Failed to create session: {str(e)}")
@router.get("/session/{session_id}/config", response_model=SessionConfigResponse)
async def get_config(session_id: str, candidate_id: str = Query(...)):
"""
Get session configuration for a candidate.
Args:
session_id: Session identifier (for logging)
candidate_id: Candidate UUID
Returns:
Session configuration including instructions and questions
"""
try:
config = get_session_config(candidate_id)
logger.info(f"Retrieved config for session {session_id}, candidate {candidate_id}")
return SessionConfigResponse(
candidate_name=config["candidate_name"],
job_title=config["job_title"],
instructions=config["instructions"],
questions=config["questions"],
config=config["config"]
)
except ValueError as e:
raise HTTPException(status_code=404, detail=str(e))
except Exception as e:
logger.error(f"Error getting config: {e}", exc_info=True)
raise HTTPException(status_code=500, detail=f"Failed to get config: {str(e)}")
@router.post("/session/{session_id}/save", response_model=SaveSessionResponse)
async def save_session(session_id: str, request: SaveSessionRequest):
"""
Save audio recording and transcript for a session.
This endpoint:
1. Retrieves audio chunks from the proxy using the token
2. Combines and saves the audio file
3. Saves transcript and audio path to database
Args:
session_id: Session identifier (must match request.session_id)
request: Contains candidate_id, transcript_text, and proxy_token
Returns:
Audio file path and success message
"""
if session_id != request.session_id:
raise HTTPException(status_code=400, detail="Session ID mismatch")
try:
# Import here to avoid circular dependency
import requests
# Get proxy URL from environment
proxy_url = os.getenv("WEBSOCKET_PROXY_URL", "ws://localhost:8000/ws/realtime")
proxy_base = proxy_url.replace("ws://", "http://").replace("wss://", "https://").replace("/ws/realtime", "")
# Retrieve audio chunks from proxy
try:
response = requests.post(
f"{proxy_base}/audio/retrieve",
params={"token": request.proxy_token},
json={"session_id": session_id},
timeout=30
)
response.raise_for_status()
audio_data = response.json()
import base64
user_chunks = audio_data.get("user_chunks", [])
# Decode Base64 audio data
for chunk in user_chunks:
if isinstance(chunk.get("data"), str):
chunk["data"] = base64.b64decode(chunk["data"])
agent_chunks = audio_data.get("agent_chunks", [])
# Decode Base64 audio data
for chunk in agent_chunks:
if isinstance(chunk.get("data"), str):
chunk["data"] = base64.b64decode(chunk["data"])
session_start_time = audio_data.get("session_start_time")
# Get transcript from proxy if available
proxy_transcript = audio_data.get("transcript", [])
transcript_text = request.transcript_text
if proxy_transcript:
logger.info(f"Using transcript from proxy ({len(proxy_transcript)} entries)")
transcript_text = "\n".join([
f"{entry.get('speaker', 'unknown')}: {entry.get('text', '')}"
for entry in proxy_transcript
])
if not session_start_time:
raise ValueError("Session start time not found in proxy response")
logger.info(f"Audio Debug: Retrieved {len(user_chunks)} user chunks and {len(agent_chunks)} agent chunks")
if user_chunks:
logger.info(f"Audio Debug: First user chunk size: {len(user_chunks[0].get('data', b''))} bytes")
except Exception as e:
logger.error(f"Error retrieving audio from proxy: {e}")
raise HTTPException(status_code=500, detail=f"Failed to retrieve audio from proxy: {str(e)}")
# Combine audio chunks
audio_file_path = None
if user_chunks or agent_chunks:
try:
logger.info("Audio Debug: Combining audio chunks...")
wav_data = combine_and_export_audio(
user_chunks=user_chunks,
agent_chunks=agent_chunks,
session_start_time=session_start_time,
session_id=session_id
)
logger.info(f"Audio Debug: Generated WAV data size: {len(wav_data)} bytes")
# Save WAV file
recordings_dir = Path("src/backend/database/voice_recordings")
recordings_dir.mkdir(parents=True, exist_ok=True)
audio_file_path = str(recordings_dir / f"{session_id}.wav")
with open(audio_file_path, "wb") as f:
f.write(wav_data)
logger.info(f"Saved audio file: {audio_file_path}")
# Verify file exists and size
if os.path.exists(audio_file_path):
size = os.path.getsize(audio_file_path)
logger.info(f"Audio Debug: File verified on disk. Size: {size} bytes")
else:
logger.error("Audio Debug: File NOT found on disk after writing!")
except Exception as e:
logger.error(f"Error processing audio: {e}", exc_info=True)
# Continue even if audio fails - we still want to save the transcript
else:
logger.warning("Audio Debug: No audio chunks found to process!")
# Save to database
try:
save_voice_screening_session(
candidate_id=request.candidate_id,
session_id=session_id,
transcript_text=transcript_text,
audio_url=audio_file_path
)
except ValueError as e:
# Candidate not found
logger.warning(f"Failed to save session: {e}")
raise HTTPException(status_code=404, detail=str(e))
logger.info(f"Saved session {session_id} for candidate {request.candidate_id}")
return SaveSessionResponse(
audio_file_path=audio_file_path,
message="Session saved successfully"
)
except HTTPException:
raise
except Exception as e:
logger.error(f"Error saving session: {e}", exc_info=True)
raise HTTPException(status_code=500, detail=f"Failed to save session: {str(e)}")
@router.get("/health")
async def health_check():
"""Health check endpoint."""
return {"status": "healthy", "service": "voice-screener"}
|