Spaces:
Building
Building
File size: 18,066 Bytes
913b6ab 196c253 59008f8 7e8a8d2 1b09f48 7e8a8d2 913b6ab c3009b9 913b6ab 2d36adf 913b6ab 7e8a8d2 913b6ab 7e8a8d2 913b6ab 7e8a8d2 1300234 7e8a8d2 913b6ab 7e8a8d2 913b6ab 7e8a8d2 76c89e5 6b6cbd5 7e8a8d2 6b6cbd5 2d36adf 913b6ab 59098f2 913b6ab 1b09f48 913b6ab 7e8a8d2 913b6ab 7e8a8d2 913b6ab 7e8a8d2 f766698 913b6ab 7e8a8d2 0b76791 913b6ab 7e8a8d2 59098f2 913b6ab d95329c 913b6ab 5997f53 913b6ab 5997f53 913b6ab 5997f53 913b6ab 1b09f48 913b6ab 1b09f48 76c89e5 913b6ab 5997f53 a0b275b 1b09f48 59008f8 25ef4b4 59008f8 913b6ab 196c253 913b6ab 5997f53 913b6ab e844e8b 0d4d92c 10974b2 5d9bc30 551e7b3 3bf8e14 134a561 551e7b3 3bf8e14 134a561 551e7b3 3bf8e14 134a561 551e7b3 3bf8e14 134a561 551e7b3 3bf8e14 134a561 0d4d92c 3bf8e14 134a561 5d9bc30 551e7b3 3bf8e14 196c253 0d4d92c 913b6ab 3bf8e14 5997f53 59098f2 7e8a8d2 0af7a49 913b6ab 2365945 913b6ab 76c89e5 7e8a8d2 913b6ab 5997f53 7e8a8d2 913b6ab 76c89e5 913b6ab 7e8a8d2 913b6ab 7e8a8d2 913b6ab 5997f53 913b6ab d3cf0d7 913b6ab 899efba 913b6ab 7e8a8d2 913b6ab 7e8a8d2 c76f23c 136867d c76f23c 913b6ab 76c89e5 119016a 913b6ab 7e8a8d2 913b6ab c76f23c 59098f2 913b6ab 5997f53 913b6ab 5997f53 913b6ab 5997f53 3bf8e14 5997f53 3bf8e14 5997f53 76c89e5 913b6ab 7e8a8d2 76c89e5 913b6ab c1df335 3a8851a c1df335 913b6ab 76c89e5 913b6ab 59098f2 913b6ab 7e8a8d2 304bfcc 7e8a8d2 304bfcc 913b6ab 304bfcc 913b6ab |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 |
from flask import Flask, request, jsonify, Response, stream_with_context, render_template_string
from google.generativeai.types import BlockedPromptException, StopCandidateException, generation_types
from google.api_core.exceptions import InvalidArgument, ResourceExhausted, Aborted, InternalServerError, ServiceUnavailable, PermissionDenied
import google.generativeai as genai
import json
import os
import re
import logging
import func
from datetime import datetime, timedelta
from apscheduler.schedulers.background import BackgroundScheduler
import time
import requests
from collections import deque
import random
os.environ['TZ'] = 'Asia/Shanghai'
app = Flask(__name__)
app.secret_key = os.urandom(24)
formatter = logging.Formatter('%(message)s')
logger = logging.getLogger(__name__)
logger.setLevel(logging.INFO)
handler = logging.StreamHandler()
handler.setFormatter(formatter)
logger.addHandler(handler)
MAX_RETRIES = int(os.environ.get('MaxRetries', 3))
MAX_REQUESTS = int(os.environ.get('MaxRequests', 2))
LIMIT_WINDOW = int(os.environ.get('LimitWindow', 60))
RETRY_DELAY = 1
MAX_RETRY_DELAY = 16
request_counts = {}
api_key_blacklist = set()
api_key_blacklist_duration = 60
# 核心优势
safety_settings = [
{
"category": "HARM_CATEGORY_HARASSMENT",
"threshold": "BLOCK_NONE"
},
{
"category": "HARM_CATEGORY_HATE_SPEECH",
"threshold": "BLOCK_NONE"
},
{
"category": "HARM_CATEGORY_SEXUALLY_EXPLICIT",
"threshold": "BLOCK_NONE"
},
{
"category": "HARM_CATEGORY_DANGEROUS_CONTENT",
"threshold": "BLOCK_NONE"
},
]
class APIKeyManager:
def __init__(self):
self.api_keys = re.findall(r"AIzaSy[a-zA-Z0-9_-]{33}", os.environ.get('KeyArray'))
self.current_index = random.randint(0, len(self.api_keys) - 1)
def get_available_key(self):
num_keys = len(self.api_keys)
for _ in range(num_keys):
if self.current_index >= num_keys:
self.current_index = 0
current_key = self.api_keys[self.current_index]
self.current_index += 1
if current_key not in api_key_blacklist:
return current_key
logger.error("所有API key都已耗尽或被暂时禁用，请重新配置或稍后重试")
return None
def show_all_keys(self):
logger.info(f"当前可用API key个数: {len(self.api_keys)} ")
for i, api_key in enumerate(self.api_keys):
logger.info(f"API Key{i}: {api_key[:11]}...")
def blacklist_key(self, key):
logger.warning(f"{key[:11]} → 暂时禁用 {api_key_blacklist_duration} 秒")
api_key_blacklist.add(key)
scheduler.add_job(lambda: api_key_blacklist.discard(key), 'date', run_date=datetime.now() + timedelta(seconds=api_key_blacklist_duration))
key_manager = APIKeyManager()
key_manager.show_all_keys()
current_api_key = key_manager.get_available_key()
def switch_api_key():
global current_api_key
key = key_manager.get_available_key()
if key:
current_api_key = key
logger.info(f"API key 替换为 → {current_api_key[:11]}...")
else:
logger.error("API key 替换失败，所有API key都已耗尽或被暂时禁用，请重新配置或稍后重试")
logger.info(f"当前 API key: {current_api_key[:11]}...")
GEMINI_MODELS = [
{"id": "gemini-1.5-flash-8b-latest"},
{"id": "gemini-1.5-flash-8b-exp-0924"},
{"id": "gemini-1.5-flash-latest"},
{"id": "gemini-1.5-flash-exp-0827"},
{"id": "gemini-1.5-pro-latest"},
{"id": "gemini-1.5-pro-exp-0827"},
{"id": "learnlm-1.5-pro-experimental"},
{"id": "gemini-exp-1114"},
{"id": "gemini-exp-1121"},
{"id": "gemini-exp-1206"},
{"id": "gemini-2.0-flash-exp"},
{"id": "gemini-2.0-flash-thinking-exp-1219"},
{"id": "gemini-2.0-pro-exp"}
]
@app.route('/')
def index():
main_content = "Moonfanz Reminiproxy v2.1.4 2025-01-09"
html_template = """
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8">
<script>
function copyToClipboard(text) {
var textarea = document.createElement("textarea");
textarea.textContent = text;
textarea.style.position = "fixed";
document.body.appendChild(textarea);
textarea.select();
try {
return document.execCommand("copy");
} catch (ex) {
console.warn("Copy to clipboard failed.", ex);
return false;
} finally {
document.body.removeChild(textarea);
}
}
function copyLink(event) {
event.preventDefault();
const url = new URL(window.location.href);
const link = url.protocol + '//' + url.host + '/hf/v1';
copyToClipboard(link);
alert('链接已复制: ' + link);
}
</script>
</head>
<body>
{{ main_content }}<br/><br/>完全开源、免费且禁止商用<br/><br/>点击复制反向代理: <a href="v1" onclick="copyLink(event)">Copy Link</a><br/>聊天来源选择"自定义(兼容 OpenAI)"<br/>将复制的网址填入到自定义端点<br/>将设置password填入自定义API秘钥<br/><br/><br/>
</body>
</html>
"""
return render_template_string(html_template, main_content=main_content)
def is_within_rate_limit(api_key):
now = datetime.now()
if api_key not in request_counts:
request_counts[api_key] = deque()
while request_counts[api_key] and request_counts[api_key][0] < now - timedelta(seconds=LIMIT_WINDOW):
request_counts[api_key].popleft()
if len(request_counts[api_key]) >= MAX_REQUESTS:
earliest_request_time = request_counts[api_key][0]
wait_time = (earliest_request_time + timedelta(seconds=LIMIT_WINDOW)) - now
return False, wait_time.total_seconds()
else:
return True, 0
def increment_request_count(api_key):
now = datetime.now()
if api_key not in request_counts:
request_counts[api_key] = deque()
request_counts[api_key].append(now)
def handle_api_error(error, attempt):
if attempt > MAX_RETRIES:
logger.error(f"{MAX_RETRIES} 次尝试后仍然失败，请修改预设或输入")
return 0, jsonify({
'error': {
'message': f"{MAX_RETRIES} 次尝试后仍然失败，请修改预设或输入",
'type': 'max_retries_exceeded'
}
})
if isinstance(error, InvalidArgument):
logger.error(f"{current_api_key[:11]} → 无效，可能已过期或被删除")
key_manager.blacklist_key(current_api_key)
switch_api_key()
return 0, None
elif isinstance(error, ResourceExhausted):
delay = min(RETRY_DELAY * (2 ** attempt), MAX_RETRY_DELAY)
logger.warning(f"{current_api_key[:11]} → 429 官方资源耗尽 → {delay} 秒后重试...")
key_manager.blacklist_key(current_api_key)
switch_api_key()
time.sleep(delay)
return 0, None
elif isinstance(error, Aborted):
delay = min(RETRY_DELAY * (2 ** attempt), MAX_RETRY_DELAY)
logger.warning(f"{current_api_key[:11]} → 操作被中止 → {delay} 秒后重试...")
time.sleep(delay)
return 0, None
elif isinstance(error, InternalServerError):
delay = min(RETRY_DELAY * (2 ** attempt), MAX_RETRY_DELAY)
logger.warning(f"{current_api_key[:11]} → 500 服务器内部错误 → {delay} 秒后重试...")
time.sleep(delay)
return 0, None
elif isinstance(error, ServiceUnavailable):
delay = min(RETRY_DELAY * (2 ** attempt), MAX_RETRY_DELAY)
logger.warning(f"{current_api_key[:11]} → 503 服务不可用 → {delay} 秒后重试...")
time.sleep(delay)
return 0, None
elif isinstance(error, PermissionDenied):
logger.error(f"{current_api_key[:11]} → 403 权限被拒绝，该 API KEY 可能已经被官方封禁")
key_manager.blacklist_key(current_api_key)
switch_api_key()
return 0, None
elif isinstance(error, StopCandidateException):
logger.warning(f"AI输出内容被Gemini官方阻挡，代理没有得到有效回复")
switch_api_key()
return 0, None
elif isinstance(error, generation_types.BlockedPromptException):
try:
full_reason_str = str(error.args[0])
if "block_reason:" in full_reason_str:
start_index = full_reason_str.find("block_reason:") + len("block_reason:")
block_reason_str = full_reason_str[start_index:].strip()
if block_reason_str == "SAFETY":
logger.warning(f"用户输入因安全原因被阻止")
return 1, None
elif block_reason_str == "BLOCKLIST":
logger.warning(f"用户输入因包含阻止列表中的术语而被阻止")
return 1, None
elif block_reason_str == "PROHIBITED_CONTENT":
logger.warning(f"用户输入因包含禁止内容而被阻止")
return 1, None
elif block_reason_str == "OTHER":
logger.warning(f"用户输入因未知原因被阻止")
return 1, None
else:
logger.warning(f"用户输入被阻止，原因未知: {block_reason_str}")
return 1, None
else:
logger.warning(f"用户输入被阻止，原因未知: {full_reason_str}")
return 1, None
except (IndexError, AttributeError) as e:
logger.error(f"获取提示原因失败↙\n{e}")
logger.error(f"提示被阻止↙\n{error}")
return 2, None
else:
logger.error(f"该模型还未发布，暂时不可用，请更换模型或未来一段时间再试")
logger.error(f"证明↙\n{error}")
return 2, None
@app.route('/hf/v1/chat/completions', methods=['POST'])
def chat_completions():
is_authenticated, auth_error, status_code = func.authenticate_request(request)
if not is_authenticated:
return auth_error if auth_error else jsonify({'error': '未授权'}), status_code if status_code else 401
request_data = request.get_json()
messages = request_data.get('messages', [])
model = request_data.get('model', 'gemini-2.0-flash-exp')
temperature = request_data.get('temperature', 1)
max_tokens = request_data.get('max_tokens', 8192)
stream = request_data.get('stream', False)
hint = "流式" if stream else "非流"
logger.info(f"\n{model} [{hint}] → {current_api_key[:11]}...")
gemini_history, user_message, error_response = func.process_messages_for_gemini(messages)
if error_response:
logger.error(f"处理输入消息时出错↙\n {error_response}")
return jsonify(error_response), 400
def do_request(current_api_key, attempt):
isok, time = is_within_rate_limit(current_api_key)
if not isok:
logger.warning(f"{current_api_key[:11]} → 暂时超过限额，该API key将在 {time} 秒后启用...")
switch_api_key()
return 0, None
increment_request_count(current_api_key)
genai.configure(api_key=current_api_key)
generation_config = {
"temperature": temperature,
"max_output_tokens": max_tokens
}
gen_model = genai.GenerativeModel(
model_name=model,
generation_config=generation_config,
safety_settings=safety_settings
)
try:
if gemini_history:
chat_session = gen_model.start_chat(history=gemini_history)
response = chat_session.send_message(user_message, stream=stream)
else:
response = gen_model.generate_content(user_message, stream=stream)
return 1, response
except Exception as e:
return handle_api_error(e, attempt)
def generate(response):
try:
logger.info(f"流式开始...")
for chunk in response:
if chunk.text:
data = {
'choices': [
{
'delta': {
'content': chunk.text
},
'finish_reason': None,
'index': 0
}
],
'object': 'chat.completion.chunk'
}
yield f"data: {json.dumps(data)}\n\n"
data = {
'choices': [
{
'delta': {},
'finish_reason': 'stop',
'index': 0
}
],
'object': 'chat.completion.chunk'
}
logger.info(f"流式结束")
yield f"data: {json.dumps(data)}\n\n"
logger.info(f"200!")
except Exception:
logger.error(f"流式输出中途被截断，请关闭流式输出或修改你的输入")
logger.info(f"流式结束")
error_data = {
'error': {
'message': '流式输出时截断，请关闭流式输出或修改你的输入',
'type': 'internal_server_error'
}
}
yield f"data: {json.dumps(error_data)}\n\n"
data = {
'choices': [
{
'delta': {},
'finish_reason': 'stop',
'index': 0
}
],
'object': 'chat.completion.chunk'
}
yield f"data: {json.dumps(data)}\n\n"
attempt = 0
success = 0
response = None
for attempt in range(1, MAX_RETRIES + 1):
logger.info(f"第 {attempt}/{MAX_RETRIES} 次尝试 ...")
success, response = do_request(current_api_key, attempt)
if success == 1:
break
elif success == 2:
logger.error(f"{model} 很可能暂时不可用，请更换模型或未来一段时间再试")
response = {
'error': {
'message': f'{model} 很可能暂时不可用，请更换模型或未来一段时间再试',
'type': 'internal_server_error'
}
}
return jsonify(response), 503
else:
logger.error(f"{MAX_RETRIES} 次尝试均失败，请调整配置或向Moonfanz反馈")
response = {
'error': {
'message': f'{MAX_RETRIES} 次尝试均失败，请调整配置或向Moonfanz反馈',
'type': 'internal_server_error'
}
}
return jsonify(response), 500 if response is not None else 503
if stream:
return Response(stream_with_context(generate(response)), mimetype='text/event-stream')
else:
try:
text_content = response.text
except (AttributeError, IndexError, TypeError, ValueError) as e:
if "response.candidates is empty" in str(e):
logger.error(f"你的输入被AI安全过滤器阻止")
return jsonify({
'error': {
'message': '你的输入被AI安全过滤器阻止',
'type': 'prompt_blocked_error',
'details': str(e)
}
}), 400
else:
logger.error(f"AI响应处理失败")
return jsonify({
'error': {
'message': 'AI响应处理失败',
'type': 'response_processing_error'
}
}), 500
response_data = {
'id': 'chatcmpl-xxxxxxxxxxxx',
'object': 'chat.completion',
'created': int(datetime.now().timestamp()),
'model': model,
'choices': [{
'index': 0,
'message': {
'role': 'assistant',
'content': text_content
},
'finish_reason': 'stop'
}],
'usage': {
'prompt_tokens': 0,
'completion_tokens': 0,
'total_tokens': 0
}
}
logger.info(f"200!")
return jsonify(response_data)
@app.route('/hf/v1/models', methods=['GET'])
def list_models():
response = {"object": "list", "data": GEMINI_MODELS}
return jsonify(response)
def keep_alive():
try:
response = requests.get("http://127.0.0.1:7860/", timeout=10)
response.raise_for_status()
print(f"Keep alive ping successful: {response.status_code} at {time.ctime()}")
except requests.exceptions.RequestException as e:
print(f"Keep alive ping failed: {e} at {time.ctime()}")
if __name__ == '__main__':
scheduler = BackgroundScheduler()
scheduler.add_job(keep_alive, 'interval', hours=12)
scheduler.start()
logger.info(f"最大尝试次数/MaxRetries: {MAX_RETRIES}")
logger.info(f"最大请求次数/MaxRequests: {MAX_REQUESTS}")
logger.info(f"请求限额窗口/LimitWindow: {LIMIT_WINDOW} 秒")
app.run(debug=True, host='0.0.0.0', port=int(os.environ.get('PORT', 7860))) |