Spaces:
Sleeping
Sleeping
File size: 16,258 Bytes
319d5b0 8c27686 319d5b0 f011a54 319d5b0 f011a54 319d5b0 f011a54 319d5b0 f011a54 319d5b0 f011a54 319d5b0 f011a54 319d5b0 f011a54 319d5b0 e2a9424 319d5b0 e2a9424 319d5b0 e2a9424 319d5b0 f011a54 319d5b0 f011a54 319d5b0 f011a54 319d5b0 f011a54 e2a9424 319d5b0 94dbf04 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 |
import os
import json
import time
import httpx
import uvicorn
import gradio as gr
from fastapi import FastAPI, Request, Header, HTTPException
from fastapi.responses import JSONResponse, StreamingResponse
# ---------------------------------------------------------------------
# Configuration
# ---------------------------------------------------------------------
BYTEZ_CHAT_URL = "https://api.bytez.com/models/v2/openai/v1/chat/completions"
BYTEZ_MODELS_URL = "https://api.bytez.com/models/v2/list/models"
BYTEZ_IMAGE_URL = "https://api.bytez.com/models/v2/google/imagen-4.0-ultra-generate-001"
BYTEZ_AUTH = os.getenv("BYTEZ_API_KEY") # your Bytez key
LOCAL_API_KEY = os.getenv("LOCAL_API_KEY") # optional local guard
BYTEZ_AUTH_2 = os.getenv("BYTEZ_API_KEY_2")
# ---------------------------------------------------------------------
# FastAPI app
# ---------------------------------------------------------------------
api = FastAPI(title="Bytez → OpenAI Proxy (v1 + v2)")
def check_key(auth: str | None):
"""Validate the Bearer token (optional local key)."""
if not auth or not auth.startswith("Bearer "):
raise HTTPException(status_code=401, detail="Missing or invalid API key")
user_key = auth.split("Bearer ")[1].strip()
if LOCAL_API_KEY and user_key != LOCAL_API_KEY:
raise HTTPException(status_code=403, detail="Unauthorized API key")
# ---------------------------------------------------------------------
# Root / health
# ---------------------------------------------------------------------
@api.get("/")
def root():
return {"status": "ok", "message": "Bytez proxy (v1+v2) running"}
# ---------------------------------------------------------------------
# -------------------------- /v1 ------------------------------------
# ---------------------------------------------------------------------
@api.get("/v1/models")
async def v1_models(authorization: str = Header(None)):
check_key(authorization)
if not BYTEZ_AUTH:
raise HTTPException(status_code=500, detail="Server BYTEZ_API_KEY not configured")
async with httpx.AsyncClient(timeout=30) as c:
r = await c.get(BYTEZ_MODELS_URL, headers={"Authorization": BYTEZ_AUTH})
try:
data = r.json()
except json.JSONDecodeError:
raise HTTPException(status_code=502, detail="Upstream returned invalid JSON")
# Transform Bytez → OpenAI list
models_list = [
{"id": m.get("id") or m.get("name"), "object": "model"}
for m in (data if isinstance(data, list) else data.get("data", []))
]
return JSONResponse(
{"object": "list", "data": models_list},
headers={"Access-Control-Allow-Origin": "*"}
)
@api.post("/v1/chat/completions")
async def v1_chat(request: Request, authorization: str = Header(None)):
"""Exactly the same implementation you already had – untouched."""
check_key(authorization)
if not BYTEZ_AUTH:
raise HTTPException(status_code=500, detail="Server BYTEZ_API_KEY not configured")
payload = await request.json()
stream = payload.get("stream", False)
headers = {"Authorization": BYTEZ_AUTH, "Content-Type": "application/json"}
# ---------- streaming helper ----------
async def v1_event_stream():
async with httpx.AsyncClient(timeout=120) as client:
async with client.stream("POST", BYTEZ_CHAT_URL, headers=headers, json=payload) as upstream:
async for line in upstream.aiter_lines():
line = line.strip()
if not line:
continue
json_str = line[6:] if line.startswith("data: ") else line
try:
chunk = json.loads(json_str)
except json.JSONDecodeError:
continue
if json_str == "[DONE]":
yield "data: [DONE]\n\n"
break
# ----- adapt Bytez chunk to OpenAI -----
content = ""
if "token" in chunk:
content = chunk["token"]
elif "choices" in chunk and chunk["choices"]:
delta = chunk["choices"][0].get("delta", {})
content = delta.get("content", "")
elif "text" in chunk:
content = chunk["text"]
else:
content = str(chunk)
openai_chunk = {
"id": "chatcmpl-proxy-stream",
"object": "chat.completion.chunk",
"created": int(time.time()),
"model": payload.get("model", "unknown"),
"choices": [
{
"index": 0,
"delta": {"role": "assistant", "content": content},
"finish_reason": None,
}
],
}
yield f"data: {json.dumps(openai_chunk)}\n\n"
yield "data: [DONE]\n\n"
# ---------- non-stream ----------
if not stream:
async with httpx.AsyncClient(timeout=120) as c:
r = await c.post(BYTEZ_CHAT_URL, headers=headers, json=payload)
try:
data = r.json()
except json.JSONDecodeError:
raise HTTPException(status_code=502, detail="Upstream returned invalid JSON")
if "choices" not in data:
content = data.get("output") or data.get("response") or data.get("message") or str(data)
data = {
"id": "chatcmpl-proxy",
"object": "chat.completion",
"choices": [{"index": 0, "message": {"role": "assistant", "content": content}}],
}
return JSONResponse(data, headers={"Access-Control-Allow-Origin": "*"})
return StreamingResponse(
v1_event_stream(),
media_type="text/event-stream",
headers={"Access-Control-Allow-Origin": "*"},
)
# ---------------------------------------------------------------------
# --------------------- /v1/images/generations (FIXED) ---------------
# ---------------------------------------------------------------------
# ---------------------------------------------------------------------
# --------------------- /v1/images/generations -----------------------
# ---------------------------------------------------------------------
@api.post("/v1/images/generations")
async def v1_images_generations(request: Request, authorization: str = Header(None)):
"""
OpenAI-compatible image generation endpoint → Bytez DALL-E 3
"""
check_key(authorization)
if not BYTEZ_AUTH:
raise HTTPException(status_code=500, detail="Server BYTEZ_API_KEY not configured")
# -------- Parse request --------
try:
payload = await request.json()
except json.JSONDecodeError:
raise HTTPException(status_code=400, detail="Invalid JSON")
prompt = payload.get("prompt")
if not prompt:
raise HTTPException(status_code=400, detail="Field 'prompt' is required")
n = int(payload.get("n", 1))
response_format = payload.get("response_format", "url") # url | b64_json
# -------- Bytez request format --------
bytez_payload = {
"text": prompt
}
headers = {
"Authorization": f"Bearer {BYTEZ_AUTH}",
"Content-Type": "application/json",
}
# -------- Call Bytez --------
async with httpx.AsyncClient(timeout=200) as client:
try:
resp = await client.post(BYTEZ_IMAGE_URL, json=bytez_payload, headers=headers)
resp.raise_for_status()
except httpx.HTTPStatusError as e:
detail = None
try:
detail = e.response.json()
except:
detail = e.response.text
raise HTTPException(status_code=e.response.status_code, detail={"upstream_error": detail})
try:
bytez_data = resp.json()
except:
raise HTTPException(status_code=502, detail="Bytez returned invalid JSON")
# -----------------------------------------------------------------
# -------- Extract URL (bytez_data.output) --------
# -----------------------------------------------------------------
url_output = None
if isinstance(bytez_data, dict):
url_output = bytez_data.get("output")
# -----------------------------------------------------------------
# -------- Extract Base64 (provider.generatedImages[].imageBytes) --
# -----------------------------------------------------------------
b64_list = []
try:
generated = bytez_data["provider"]["generatedImages"]
for g in generated:
img = g.get("image", {})
if "imageBytes" in img:
b64_list.append(img["imageBytes"])
except Exception:
pass # ignore if provider/genImages is missing
# -----------------------------------------------------------------
# -------- Build OpenAI-compatible response ------------------------
# -----------------------------------------------------------------
openai_data = []
count = max(len(b64_list), 1)
count = min(count, n)
for i in range(count):
# --- Base64 Output ---
if response_format == "b64_json":
if i < len(b64_list):
openai_data.append({"b64_json": b64_list[i]})
else:
# fallback: turn URL into base64-json payload (OpenAI accepts raw string)
openai_data.append({"b64_json": url_output})
else:
# --- URL Output ---
if url_output:
openai_data.append({"url": url_output})
elif i < len(b64_list):
# fallback: convert base64 into a data URL
openai_data.append({"url": f"data:image/png;base64,{b64_list[i]}"})
else:
openai_data.append({"url": ""})
result = {
"created": int(time.time()),
"data": openai_data,
}
return JSONResponse(result, headers={"Access-Control-Allow-Origin": "*"})
# ---------------------------------------------------------------------
# -------------------------- /v2 ------------------------------------
# ---------------------------------------------------------------------
@api.post("/v2/chat/completions")
async def v2_chat_completions(request: Request, authorization: str = Header(None)):
"""
v2 – clean OpenAI-compatible streaming.
* First chunk includes role=assistant (required by Continue.dev)
* Later chunks send only delta.content
* No usage events
"""
check_key(authorization)
if not BYTEZ_AUTH_2:
raise HTTPException(status_code=500, detail="Server BYTEZ_API_2 not configured")
try:
body = await request.body()
payload = json.loads(body.decode("utf-8"))
except json.JSONDecodeError as e:
raise HTTPException(status_code=400, detail=f"Invalid JSON: {e}")
stream = payload.get("stream", False)
upstream_headers = {
"Authorization": BYTEZ_AUTH_2,
"Content-Type": "application/json",
}
# Normal content chunk (NO ROLE)
def make_openai_delta(content: str):
return {
"id": f"chatcmpl-v2-{int(time.time())}",
"object": "chat.completion.chunk",
"created": int(time.time()),
"model": payload.get("model", "unknown"),
"choices": [
{
"index": 0,
"delta": {"content": content},
"finish_reason": None,
}
],
}
async def clean_stream():
# FIRST CHUNK MUST SET THE ROLE → REQUIRED by Continue.dev
first_chunk = {
"id": f"chatcmpl-v2-{int(time.time())}",
"object": "chat.completion.chunk",
"created": int(time.time()),
"model": payload.get("model", "unknown"),
"choices": [
{
"index": 0,
"delta": {"role": "assistant", "content": ""},
"finish_reason": None,
}
],
}
# Send first role-setting chunk
yield f"data: {json.dumps(first_chunk)}\n\n"
async with httpx.AsyncClient(timeout=180) as client:
try:
async with client.stream(
"POST", BYTEZ_CHAT_URL, headers=upstream_headers, json=payload
) as upstream:
async for line in upstream.aiter_lines():
line = line.strip()
if not line:
continue
json_str = line[6:] if line.startswith("data: ") else line
# Skip usage events
if "usage" in json_str.lower():
continue
if json_str == "[DONE]":
yield "data: [DONE]\n\n"
return
try:
chunk = json.loads(json_str)
except json.JSONDecodeError:
continue
text = ""
if isinstance(chunk, dict):
if "token" in chunk:
text = chunk["token"]
elif "choices" in chunk and chunk["choices"]:
delta = chunk["choices"][0].get("delta", {})
text = delta.get("content", "")
elif "text" in chunk:
text = chunk["text"]
else:
text = str(chunk)
if text:
yield f"data: {json.dumps(make_openai_delta(text))}\n\n"
yield "data: [DONE]\n\n"
except Exception as e:
error_chunk = make_openai_delta(f"Error: {str(e)}")
yield f"data: {json.dumps(error_chunk)}\n\n"
yield "data: [DONE]\n\n"
# Non-streaming mode
if not stream:
async with httpx.AsyncClient(timeout=120) as c:
r = await c.post(BYTEZ_CHAT_URL, headers=upstream_headers, json=payload)
r.raise_for_status()
data = r.json()
if "choices" not in data:
content = (
data.get("output")
or data.get("response")
or data.get("message")
or str(data)
)
data = {
"id": "chatcmpl-v2",
"object": "chat.completion",
"choices": [
{"index": 0, "message": {"role": "assistant", "content": content}}
],
}
return JSONResponse(data, headers={"Access-Control-Allow-Origin": "*"})
# Streaming mode
return StreamingResponse(
clean_stream(),
media_type="text/event-stream",
headers={
"Access-Control-Allow-Origin": "*",
"Access-Control-Allow-Headers": "*",
"Cache-Control": "no-cache",
"Connection": "keep-alive",
"X-Accel-Buffering": "no",
},
)
# ---------------------------------------------------------------------
# Minimal Gradio UI (required for HF Space to start)
# ---------------------------------------------------------------------
with gr.Blocks() as ui:
gr.Markdown(
"### Bytez → OpenAI Proxy (v1 + **v2**)\n"
"- `/v1/models` \n"
"- `/v1/chat/completions` (unchanged) \n"
"- **`/v2/chat/completions`** – clean streaming, no usage chunk"
)
demo = gr.mount_gradio_app(api, ui, path="/")
# This makes it work on Render, Railway, Fly.io, etc.
app = api
if __name__ == "__main__":
# Only for local testing with Gradio
uvicorn.run(demo, host="0.0.0.0", port=7860)
|