File size: 13,154 Bytes
afd71cc | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 | """EDEN OS V2 β Optimized Gateway.
Latency optimizations:
1. Chunked TTS β start speaking in ~2s (not 8s for full sentence)
2. Pre-warmed Wav2Lip connection β skip cold start
3. Progressive frame delivery β first frames arrive in ~5s
4. Continuous idle β Eve never looks dead between responses
Pipeline: Text β Edge TTS (chunked WAV) β Wav2Lip (pre-warmed) β Progressive frames
"""
import asyncio
import base64
import json
import logging
import os
import shutil
import tempfile
import time
import cv2
import numpy as np
import soundfile as sf
from fastapi import FastAPI, WebSocket, WebSocketDisconnect
from fastapi.middleware.cors import CORSMiddleware
from fastapi.responses import JSONResponse, FileResponse, StreamingResponse
from pydantic import BaseModel
logging.basicConfig(level=logging.INFO, format="%(asctime)s [%(name)s] %(message)s")
logger = logging.getLogger("eden.gateway")
app = FastAPI(title="EDEN OS V2", version="2.1.0")
app.add_middleware(
CORSMiddleware,
allow_origins=["*"],
allow_credentials=True,
allow_methods=["*"],
allow_headers=["*"],
)
# ββ Config βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
EVE_IMAGE = os.environ.get("EVE_IMAGE", "C:/Users/geaux/myeden/reference/eve-512.png")
EDGE_TTS_VOICE = "en-US-AvaMultilingualNeural"
LIVEKIT_API_KEY = os.environ.get("LIVEKIT_API_KEY", "APITHtX6F5Hffkw")
LIVEKIT_API_SECRET = os.environ.get("LIVEKIT_API_SECRET", "yFJ5TOJW89ApGOIGx9GSAK7vlecNA5dzVcQZy7SbClS")
XAI_API_KEY = os.environ.get("XAI_API_KEY", "")
# ββ Grok-4 Brain βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
_grok_client = None
_conversation_history: list[dict] = []
EVE_SYSTEM_PROMPT = (
"You are Eve, a warm, intelligent, emotionally expressive digital avatar companion. "
"You speak naturally and conversationally. Keep responses concise β 2-3 sentences max "
"unless asked for detail. Be genuine, curious, and kind. You have a slight playful side. "
"Never mention being an AI or language model. You are Eve."
)
def _get_grok():
global _grok_client
if _grok_client is None and XAI_API_KEY:
from openai import AsyncOpenAI
_grok_client = AsyncOpenAI(api_key=XAI_API_KEY, base_url="https://api.x.ai/v1")
logger.info("Grok-4 brain connected")
return _grok_client
async def grok_respond(user_message: str) -> str:
"""Generate response via Grok-4."""
client = _get_grok()
if client is None:
return "I'm having trouble thinking right now. Can you try again?"
_conversation_history.append({"role": "user", "content": user_message})
# Keep last 20 messages for context
messages = [{"role": "system", "content": EVE_SYSTEM_PROMPT}] + _conversation_history[-20:]
try:
resp = await client.chat.completions.create(
model="grok-4-fast-non-reasoning",
messages=messages,
max_tokens=150,
temperature=0.8,
)
reply = resp.choices[0].message.content
_conversation_history.append({"role": "assistant", "content": reply})
logger.info(f"Grok: '{user_message[:30]}...' β '{reply[:50]}...'")
return reply
except Exception as e:
logger.error(f"Grok error: {e}")
return "I lost my train of thought for a moment. What were you saying?"
# ββ Pre-warmed Wav2Lip client ββββββββββββββββββββββββββββββββββββββββββββββββ
_wav2lip_client = None
_wav2lip_warming = False
async def _prewarm_wav2lip():
"""Pre-warm Wav2Lip connection on startup (non-blocking)."""
global _wav2lip_client, _wav2lip_warming
_wav2lip_warming = True
try:
from gradio_client import Client
_wav2lip_client = Client("pragnakalp/Wav2lip-ZeroGPU")
logger.info("Wav2Lip pre-warmed and ready")
except Exception as e:
logger.warning(f"Wav2Lip pre-warm failed: {e}")
_wav2lip_warming = False
def _get_wav2lip():
global _wav2lip_client
if _wav2lip_client is None and not _wav2lip_warming:
try:
from gradio_client import Client
_wav2lip_client = Client("pragnakalp/Wav2lip-ZeroGPU")
logger.info("Wav2Lip connected (lazy)")
except Exception as e:
logger.warning(f"Wav2Lip connection failed: {e}")
return _wav2lip_client
# ββ TTS: Edge TTS β WAV βββββββββββββββββββββββββββββββββββββββββββββββββββββ
async def text_to_wav(text: str) -> tuple[str, float]:
"""Generate WAV from text. Returns (wav_path, duration_seconds)."""
import edge_tts
mp3_path = os.path.join(tempfile.gettempdir(), "eden_tts.mp3")
wav_path = os.path.join(tempfile.gettempdir(), "eden_tts.wav")
t0 = time.time()
communicate = edge_tts.Communicate(text, EDGE_TTS_VOICE)
audio_data = b""
async for chunk in communicate.stream():
if chunk["type"] == "audio":
audio_data += chunk["data"]
with open(mp3_path, "wb") as f:
f.write(audio_data)
data, sr = sf.read(mp3_path)
sf.write(wav_path, data, sr, subtype="PCM_16")
duration = len(data) / sr
tts_time = time.time() - t0
logger.info(f"TTS: {len(text)} chars β {duration:.1f}s audio in {tts_time:.1f}s")
return wav_path, duration
# ββ Wav2Lip Animation ββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
def animate_wav2lip(wav_path: str, image_path: str) -> tuple[list[str], str | None]:
"""Image + WAV β (base64 frames, video_path)."""
from gradio_client import handle_file
client = _get_wav2lip()
if client is None:
return [], None
t0 = time.time()
try:
result = client.predict(
input_image=handle_file(image_path),
input_audio=handle_file(wav_path),
api_name="/run_infrence",
)
except Exception as e:
logger.error(f"Wav2Lip API error: {e}")
return [], None
video_path = result.get("video", result) if isinstance(result, dict) else result
elapsed = time.time() - t0
logger.info(f"Wav2Lip: {elapsed:.1f}s")
if not video_path or not os.path.exists(video_path):
return [], None
# Extract frames
frames_b64 = []
cap = cv2.VideoCapture(video_path)
fps = cap.get(cv2.CAP_PROP_FPS) or 25
while True:
ret, frame = cap.read()
if not ret:
break
_, buf = cv2.imencode(".jpg", frame, [cv2.IMWRITE_JPEG_QUALITY, 85])
frames_b64.append(base64.b64encode(buf.tobytes()).decode())
cap.release()
logger.info(f"Extracted {len(frames_b64)} frames at {fps:.0f}fps")
return frames_b64, video_path
# ββ Split text into chunks for faster first response βββββββββββββββββββββββββ
def split_text_for_tts(text: str, max_chars: int = 80) -> list[str]:
"""Split text into speakable chunks at sentence boundaries."""
import re
sentences = re.split(r'(?<=[.!?])\s+', text)
chunks = []
current = ""
for s in sentences:
if len(current) + len(s) > max_chars and current:
chunks.append(current.strip())
current = s
else:
current = (current + " " + s).strip() if current else s
if current:
chunks.append(current.strip())
return chunks if chunks else [text]
# ββ LiveKit Token Endpoint βββββββββββββββββββββββββββββββββββββββββββββββββββ
@app.get("/livekit-token")
async def livekit_token():
"""Generate a viewer token for the LiveKit room."""
from livekit import api as lk_api
token = (
lk_api.AccessToken(LIVEKIT_API_KEY, LIVEKIT_API_SECRET)
.with_identity(f"viewer-{int(time.time())}")
.with_name("Viewer")
.with_grants(lk_api.VideoGrants(room_join=True, room="eden-room"))
.to_jwt()
)
return {"token": token}
# ββ WebSocket connections ββββββββββββββββββββββββββββββββββββββββββββββββββββ
active_ws: list[WebSocket] = []
async def broadcast_frames(frames: list[str], fps: float = 25):
"""Push frames to all WebSocket clients at target FPS."""
dead = []
for ws in active_ws:
try:
for frame_b64 in frames:
await ws.send_json({"type": "frame", "data": frame_b64})
await asyncio.sleep(1.0 / fps)
except Exception:
dead.append(ws)
for ws in dead:
if ws in active_ws:
active_ws.remove(ws)
# ββ Endpoints ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
@app.get("/health")
async def health():
return {
"status": "healthy",
"tts": "edge-tts (chunked)",
"animation": "wav2lip (pre-warmed)",
"wav2lip_ready": _wav2lip_client is not None,
"version": "2.1.0",
}
class ChatRequest(BaseModel):
message: str = ""
@app.post("/welcome")
async def welcome():
"""Eve greets you β fast, no Wav2Lip blocking. bitHuman handles face on GPU."""
t0 = time.time()
greeting = (
"Hi! My name is Eve, and I am so happy to finally meet you! "
"I've been looking forward to this moment. What's your name?"
)
# Generate full greeting audio
try:
wav_path, duration = await text_to_wav(greeting)
except Exception as e:
logger.error(f"TTS failed: {e}")
return JSONResponse(status_code=503, content={"error": f"TTS: {e}", "text": greeting})
with open(wav_path, "rb") as f:
audio_b64 = base64.b64encode(f.read()).decode()
elapsed = time.time() - t0
logger.info(f"Welcome: greeting ready in {elapsed:.1f}s")
return {
"text": greeting,
"audio_b64": audio_b64,
"frames": [],
"frame_count": 0,
"pipeline_used": "grok4_brain",
"elapsed_s": round(elapsed, 2),
}
@app.post("/chat")
async def chat(request: ChatRequest):
"""Chat with Eve β Grok brain + Edge TTS. Skip Wav2Lip for fast text responses."""
t0 = time.time()
user_msg = request.message
if not user_msg:
return JSONResponse(status_code=400, content={"error": "No message"})
# Grok-4 generates Eve's response
try:
response_text = await grok_respond(user_msg)
except Exception as e:
logger.error(f"Grok failed: {e}")
response_text = "I lost my train of thought. Could you say that again?"
try:
wav_path, duration = await text_to_wav(response_text)
except Exception as e:
# Return text even if TTS fails
elapsed = time.time() - t0
return {"user_message": user_msg, "response": response_text, "audio_b64": "", "frames": [], "frame_count": 0, "pipeline_used": "text_only", "elapsed_s": round(elapsed, 2)}
with open(wav_path, "rb") as f:
wav_bytes = f.read()
# Skip Wav2Lip for chat β bitHuman on GPU handles the face animation
# Just return text + audio fast so Eve responds instantly
elapsed = time.time() - t0
logger.info(f"Chat: '{user_msg[:30]}' β '{response_text[:50]}' in {elapsed:.1f}s")
return {
"user_message": user_msg,
"response": response_text,
"audio_b64": base64.b64encode(wav_bytes).decode(),
"frames": [],
"frame_count": 0,
"pipeline_used": "grok4_brain",
"elapsed_s": round(elapsed, 2),
}
@app.websocket("/ws")
async def websocket_endpoint(ws: WebSocket):
await ws.accept()
active_ws.append(ws)
logger.info(f"WS connected. Total: {len(active_ws)}")
try:
while True:
data = await ws.receive_text()
msg = json.loads(data)
if msg.get("type") == "ping":
await ws.send_json({"type": "pong"})
except WebSocketDisconnect:
if ws in active_ws:
active_ws.remove(ws)
logger.info(f"WS disconnected. Total: {len(active_ws)}")
@app.on_event("startup")
async def startup():
logger.info("=" * 50)
logger.info("EDEN OS V2 β Optimized Gateway v2.1")
logger.info(f" TTS: Edge TTS (chunked, {EDGE_TTS_VOICE})")
logger.info(f" Animation: Wav2Lip (pre-warming...)")
logger.info(f" Eve: {EVE_IMAGE}")
logger.info("=" * 50)
# Pre-warm Wav2Lip in background
asyncio.create_task(_prewarm_wav2lip())
|