返回提交历史
Modified
g4f/Provider/Qwen.py
+181
-41
Modified
g4f/Provider/Yupp.py
+190
-72
Modified
g4f/Provider/needs_auth/LMArena.py
+1
-1
Modified
g4f/Provider/needs_auth/OpenaiChat.py
+234
-171
Modified
g4f/image/__init__.py
+9
-0
Modified
g4f/requests/aiohttp.py
+8
-4
XFEstudio/gpt4free
Improve Yupp provider account handling , request timeout and get byte from url (#3249)
* Add image caching to Yupp provider Introduces an image cache to avoid redundant uploads in the Yupp provider. Refactors media attachment handling into a new prepare_files method, improving efficiency and code organization. Updates .gitignore to exclude .idea directory. * Refactor Yupp stream handling and chunk processing Improves stream segmentation in the Yupp provider by introducing buffers for target, variant, quick, thinking, and extra streams. Refactors chunk processing to better handle image-gen, quick responses, and variant outputs, and adds more robust stream ID extraction and routing logic. Yields a consolidated JsonResponse with all stream segments for downstream use. * Handle ClientResponseError in Yupp provider Adds specific handling for aiohttp ClientResponseError in the Yupp provider. Marks account as invalid on 500 Internal Server Error, otherwise increments error count and raises ProviderException for other errors. * Update Yupp.py fix 429 'Too Many Requests' * Update Yupp.py * Improve Yupp provider account handling and request timeout Refactored account loading to preserve account history and error counts when updating tokens. Enhanced request logic to support custom timeouts using aiohttp's ClientTimeout, allowing for more flexible timeout configuration. * Update __init__.py * Handle multi-line <think> and <yapp> blocks in Yupp Added logic to capture and process multi-line <think> and <yapp class="image-gen"> blocks referenced by special IDs. Introduced block storage and extraction functions, enabling reasoning and image-gen content to be handled via references in the response stream. * Update LMArena.py Not Found Model error * Refactor to use StreamSession in Qwen and Yupp providers Replaced aiohttp.ClientSession with StreamSession in Qwen.py and Yupp.py for improved session handling. Updated exception and timeout references in Yupp.py to use aiohttp types. Improved default argument handling in StreamSession initialization. * Update Yupp.py * Add status parameter to get_generated_image method Introduces a 'status' parameter to the get_generated_image method to allow passing image generation status. Updates method calls and response objects to include status in their metadata for improved tracking of image generation progress. * Update OpenaiChat.py * Refactor Qwen image upload and caching logic and token Reworked the image upload flow in Qwen provider to use direct file uploads with OSS headers, added caching for uploaded images, and improved file type detection. Updated prepare_files to handle uploads via session and cache results, and added utility for generating OSS headers. Minor imports and typing adjustments included and token support. * Refactor Qwen and Yupp providers for improved async handling Updated Qwen provider to handle timeout via kwargs and improved type annotations. Refactored Yupp provider for better code organization, formatting, and async account rotation logic. Enhanced readability and maintainability by reordering imports, adding whitespace, and clarifying function implementations. * Add image caching to OpenaiChat provider Introduces an image cache mechanism to OpenaiChat for uploaded images, reducing redundant uploads and improving efficiency. Also refactors code for clarity, updates type hints, and makes minor formatting improvements throughout the file.
7771cf3d
代码差异
6 个文件
+623
-289
@@ -1,22 +1,29 @@
1
1
from __future__ import annotations
2
2
3
3
import asyncio
4
import datetime
5
import hashlib
6
import hmac
4
7
import json
5
import mimetypes
6
8
import re
7
9
import uuid
8
10
from time import time
9
from typing import Literal, Optional
11
from typing import Literal, Optional, Dict
12
from urllib.parse import quote
10
13
11
14
import aiohttp
15
16
from g4f.image import to_bytes, detect_file_type
17
from g4f.requests import raise_for_status
18
from .base_provider import AsyncGeneratorProvider, ProviderModelMixin
19
from .helper import get_last_user_message
20
from .. import debug
12
21
from ..errors import RateLimitError, ResponseError
13
from ..typing import AsyncResult, Messages, MediaListType
14
22
from ..providers.response import JsonConversation, Reasoning, Usage, ImageResponse, FinishReason
15
23
from ..requests import sse_stream
24
from ..requests.aiohttp import StreamSession
16
25
from ..tools.media import merge_media
17
from .base_provider import AsyncGeneratorProvider, ProviderModelMixin
18
from .helper import get_last_user_message
19
from .. import debug
26
from ..typing import AsyncResult, Messages, MediaListType
20
27
21
28
try:
22
29
import curl_cffi
@@ -25,6 +32,56 @@ try:
25
32
except ImportError:
26
33
has_curl_cffi = False
27
34
35
# Global variables to manage Qwen Image Cache
36
ImagesCache: Dict[str, dict] = {}
37
38
39
def get_oss_headers(method: str, date_str: str, sts_data: dict, content_type: str) -> dict[str, str]:
40
bucket_name = sts_data.get('bucketname', 'qwen-webui-prod')
41
file_path = sts_data.get('file_path', '')
42
access_key_id = sts_data.get('access_key_id')
43
access_key_secret = sts_data.get('access_key_secret')
44
security_token = sts_data.get('security_token')
45
headers = {
46
'Content-Type': content_type,
47
'x-oss-content-sha256': 'UNSIGNED-PAYLOAD',
48
'x-oss-date': date_str,
49
'x-oss-security-token': security_token,
50
'x-oss-user-agent': 'aliyun-sdk-js/6.23.0 Chrome 132.0.0.0 on Windows 10 64-bit'
51
}
52
headers_lower = {k.lower(): v for k, v in headers.items()}
53
54
canonical_headers_list = []
55
signed_headers_list = []
56
required_headers = ['content-md5', 'content-type', 'x-oss-content-sha256', 'x-oss-date', 'x-oss-security-token',
57
'x-oss-user-agent']
58
for header_name in sorted(required_headers):
59
if header_name in headers_lower:
60
canonical_headers_list.append(f"{header_name}:{headers_lower[header_name]}")
61
signed_headers_list.append(header_name)
62
63
canonical_headers = '\n'.join(canonical_headers_list) + '\n'
64
canonical_uri = f"/{bucket_name}/{quote(file_path, safe='/')}"
65
66
canonical_request = f"{method}\n{canonical_uri}\n\n{canonical_headers}\n\nUNSIGNED-PAYLOAD"
67
68
date_parts = date_str.split('T')
69
date_scope = f"{date_parts[0]}/ap-southeast-1/oss/aliyun_v4_request"
70
string_to_sign = f"OSS4-HMAC-SHA256\n{date_str}\n{date_scope}\n{hashlib.sha256(canonical_request.encode()).hexdigest()}"
71
72
def sign(key, msg):
73
return hmac.new(key, msg.encode() if isinstance(msg, str) else msg, hashlib.sha256).digest()
74
75
date_key = sign(f"aliyun_v4{access_key_secret}".encode(), date_parts[0])
76
region_key = sign(date_key, "ap-southeast-1")
77
service_key = sign(region_key, "oss")
78
signing_key = sign(service_key, "aliyun_v4_request")
79
signature = hmac.new(signing_key, string_to_sign.encode(), hashlib.sha256).hexdigest()
80
81
headers['authorization'] = f"OSS4-HMAC-SHA256 Credential={access_key_id}/{date_scope},Signature={signature}"
82
return headers
83
84
28
85
text_models = [
29
86
'qwen3-max-preview', 'qwen-plus-2025-09-11', 'qwen3-235b-a22b', 'qwen3-coder-plus', 'qwen3-30b-a3b',
30
87
'qwen3-coder-30b-a3b-instruct', 'qwen-max-latest', 'qwen-plus-2025-01-25', 'qwq-32b', 'qwen-turbo-2025-02-11',
@@ -60,19 +117,19 @@ class Qwen(AsyncGeneratorProvider, ProviderModelMixin):
60
117
active_by_default = True
61
118
supports_stream = True
62
119
supports_message_history = False
63
120
image_cache = True
64
121
_models_loaded = True
65
122
image_models = image_models
66
123
text_models = text_models
67
124
vision_models = vision_models
68
models = models
125
models: list[str] = models
69
126
default_model = "qwen3-235b-a22b"
70
127
71
128
_midtoken: str = None
72
129
_midtoken_uses: int = 0
73
130
74
131
@classmethod
75
def get_models(cls) -> list[str]:
132
def get_models(cls, **kwargs) -> list[str]:
76
133
if not cls._models_loaded and has_curl_cffi:
77
134
response = curl_cffi.get(f"{cls.url}/api/models")
78
135
if response.ok:
@@ -97,34 +154,106 @@ class Qwen(AsyncGeneratorProvider, ProviderModelMixin):
97
154
return cls.models
98
155
99
156
@classmethod
100
async def prepare_files(cls, media, chat_type="")->list:
157
async def prepare_files(cls, media, session: aiohttp.ClientSession, headers=None) -> list:
158
if headers is None:
159
headers = {}
101
160
files = []
102
for _file, file_name in media:
103
file_type, _ = mimetypes.guess_type(file_name)
104
file_class: Literal["default", "vision", "video", "audio", "document"] = "default"
105
_type: Literal["file", "image", "video", "audio"] = "file"
106
showType: Literal["file", "image", "video", "audio"] = "file"
107
108
if isinstance(_file, str) and _file.startswith('http'):
109
if chat_type == "image_edit" or (file_type and file_type.startswith("image")):
110
file_class = "vision"
111
_type = "image"
112
if not file_type:
113
# Try to infer from file extension, fallback to generic
114
ext = file_name.split('.')[-1].lower() if '.' in file_name else ''
115
file_type = mimetypes.types_map.get(f'.{ext}', 'application/octet-stream')
116
showType = "image"
117
118
files.append(
119
{
120
"type": _type,
161
for index, (_file, file_name) in enumerate(media):
162
163
data_bytes = to_bytes(_file)
164
# Check Cache
165
hasher = hashlib.md5()
166
hasher.update(data_bytes)
167
image_hash = hasher.hexdigest()
168
file = ImagesCache.get(image_hash)
169
if cls.image_cache and file:
170
debug.log("Using cached image")
171
files.append(file)
172
continue
173
174
extension, file_type = detect_file_type(data_bytes)
175
file_name = file_name or f"file-{len(data_bytes)}{extension}"
176
file_size = len(data_bytes)
177
178
# Get File Url
179
async with session.post(
180
f'{cls.url}/api/v2/files/getstsToken',
181
json={"filename": file_name,
182
"filesize": file_size, "filetype": file_type},
183
headers=headers
184
185
) as r:
186
await raise_for_status(r, "Create file failed")
187
res_data = await r.json()
188
data = res_data.get("data")
189
190
if res_data["success"] is False:
191
raise RateLimitError(f"{data['code']}:{data['details']}")
192
file_url = data.get("file_url")
193
file_id = data.get("file_id")
194
195
# Put File into Url
196
str_date = datetime.datetime.now(datetime.UTC).strftime('%Y%m%dT%H%M%SZ')
197
headers = get_oss_headers('PUT', str_date, data, file_type)
198
async with session.put(
199
file_url.split("?")[0],
200
data=data_bytes,
201
headers=headers
202
) as response:
203
await raise_for_status(response)
204
205
file_class: Literal["default", "vision", "video", "audio", "document"]
206
_type: Literal["file", "image", "video", "audio"]
207
show_type: Literal["file", "image", "video", "audio"]
208
if "image" in file_type:
209
_type = "image"
210
show_type = "image"
211
file_class = "vision"
212
elif "video" in file_type:
213
_type = "video"
214
show_type = "video"
215
file_class = "video"
216
elif "audio" in file_type:
217
_type = "audio"
218
show_type = "audio"
219
file_class = "audio"
220
else:
221
_type = "file"
222
show_type = "file"
223
file_class = "document"
224
225
file = {
226
"type": _type,
227
"file": {
228
"created_at": int(time() * 1000),
229
"data": {},
230
"filename": file_name,
231
"hash": None,
232
"id": file_id,
233
"meta": {
121
234
"name": file_name,
122
"file_type": file_type,
123
"showType": showType,
124
"file_class": file_class,
125
"url": _file
126
}
127
)
235
"size": file_size,
236
"content_type": file_type
237
},
238
"update_at": int(time() * 1000),
239
},
240
"id": file_id,
241
"url": file_url,
242
"name": file_name,
243
"collection_name": "",
244
"progress": 0,
245
"status": "uploaded",
246
"greenNet": "success",
247
"size": file_size,
248
"error": "",
249
"itemId": str(uuid.uuid4()),
250
"file_type": file_type,
251
"showType": show_type,
252
"file_class": file_class,
253
"uploadTaskId": str(uuid.uuid4())
254
}
255
ImagesCache[image_hash] = file
256
files.append(file)
128
257
return files
129
258
130
259
@classmethod
@@ -135,7 +264,6 @@ class Qwen(AsyncGeneratorProvider, ProviderModelMixin):
135
264
media: MediaListType = None,
136
265
conversation: JsonConversation = None,
137
266
proxy: str = None,
138
timeout: int = 120,
139
267
stream: bool = True,
140
268
enable_thinking: bool = True,
141
269
chat_type: Literal[
@@ -157,7 +285,7 @@ class Qwen(AsyncGeneratorProvider, ProviderModelMixin):
157
285
"""
158
286
159
287
model_name = cls.get_model(model)
160
288
token = kwargs.get("token")
161
289
headers = {
162
290
'User-Agent': 'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/138.0.0.0 Safari/537.36',
163
291
'Accept': '*/*',
@@ -169,13 +297,24 @@ class Qwen(AsyncGeneratorProvider, ProviderModelMixin):
169
297
'Sec-Fetch-Mode': 'cors',
170
298
'Sec-Fetch-Site': 'same-origin',
171
299
'Connection': 'keep-alive',
172
'Authorization': 'Bearer',
300
'Authorization': f'Bearer {token}' if token else "Bearer",
173
301
'Source': 'web'
174
302
}
175
303
176
304
prompt = get_last_user_message(messages)
177
178
async with aiohttp.ClientSession(headers=headers) as session:
305
_timeout = kwargs.get("timeout")
306
if isinstance(_timeout, aiohttp.ClientTimeout):
307
timeout = _timeout
308
else:
309
total = float(_timeout) if isinstance(_timeout, (int, float)) else 5 * 60
310
timeout = aiohttp.ClientTimeout(total=total)
311
async with StreamSession(headers=headers) as session:
312
try:
313
async with session.get('https://chat.qwen.ai/api/v1/auths/', proxy=proxy) as user_info_res:
314
user_info_res.raise_for_status()
315
debug.log(await user_info_res.json())
316
except:
317
...
179
318
for attempt in range(5):
180
319
try:
181
320
if not cls._midtoken:
@@ -221,7 +360,8 @@ class Qwen(AsyncGeneratorProvider, ProviderModelMixin):
221
360
files = []
222
361
media = list(merge_media(media, messages))
223
362
if media:
224
files = await cls.prepare_files(media, chat_type=chat_type)
363
files = await cls.prepare_files(media, session=session,
364
headers=req_headers)
225
365
226
366
msg_payload = {
227
367
"stream": stream,
@@ -1,25 +1,26 @@
1
import asyncio
1
2
import hashlib
2
3
import json
4
import os
5
import re
3
6
import time
4
7
import uuid
5
import re
6
import os
7
import asyncio
8
8
9
import aiohttp
9
from aiohttp import ClientResponseError
10
10
11
from ..typing import AsyncResult, Messages, Optional, Dict, Any, List
11
from .helper import get_last_user_message
12
from .yupp.models import YuppModelManager
13
from ..cookies import get_cookies
14
from ..debug import log
15
from ..errors import RateLimitError, ProviderException, MissingAuthError
16
from ..image import is_accepted_format, to_bytes
12
17
from ..providers.base_provider import AsyncGeneratorProvider, ProviderModelMixin
13
18
from ..providers.response import Reasoning, PlainTextResponse, PreviewResponse, JsonConversation, ImageResponse, \
14
19
ProviderInfo, FinishReason, JsonResponse
15
from ..errors import RateLimitError, ProviderException, MissingAuthError
16
from ..cookies import get_cookies
20
from ..requests.aiohttp import StreamSession
17
21
from ..tools.auth import AuthManager
18
22
from ..tools.media import merge_media
19
from ..image import is_accepted_format, to_bytes
20
from .yupp.models import YuppModelManager
21
from .helper import get_last_user_message
22
from ..debug import log
23
from ..typing import AsyncResult, Messages, Optional, Dict, Any, List
23
24
24
25
# Global variables to manage Yupp accounts
25
26
YUPP_ACCOUNT = Dict[str, Any]
@@ -27,22 +28,24 @@ YUPP_ACCOUNTS: List[YUPP_ACCOUNT] = []
27
28
account_rotation_lock = asyncio.Lock()
28
29
29
30
# Global variables to manage Yupp Image Cache
30
ImagesCache:Dict[str, dict] = {}
31
ImagesCache: Dict[str, dict] = {}
32
31
33
32
34
class YuppAccount:
33
35
"""Yupp account representation"""
36
34
37
def __init__(self, token: str, is_valid: bool = True, error_count: int = 0, last_used: float = 0):
35
38
self.token = token
36
39
self.is_valid = is_valid
37
40
self.error_count = error_count
38
41
self.last_used = last_used
39
42
43
40
44
def load_yupp_accounts(tokens_str: str):
41
45
"""Load Yupp accounts from token string"""
42
46
global YUPP_ACCOUNTS
43
47
if not tokens_str:
44
48
return
45
46
49
tokens = [token.strip() for token in tokens_str.split(',') if token.strip()]
47
50
YUPP_ACCOUNTS = [
48
51
{
@@ -54,6 +57,7 @@ def load_yupp_accounts(tokens_str: str):
54
57
for token in tokens
55
58
]
56
59
60
57
61
def create_headers() -> Dict[str, str]:
58
62
"""Create headers for requests"""
59
63
return {
@@ -66,6 +70,7 @@ def create_headers() -> Dict[str, str]:
66
70
"Sec-Fetch-Site": "same-origin",
67
71
}
68
72
73
69
74
async def get_best_yupp_account() -> Optional[YUPP_ACCOUNT]:
70
75
"""Get the best available Yupp account using smart selection algorithm"""
71
76
max_error_count = int(os.getenv("MAX_ERROR_COUNT", "3"))
@@ -77,10 +82,10 @@ async def get_best_yupp_account() -> Optional[YUPP_ACCOUNT]:
77
82
acc
78
83
for acc in YUPP_ACCOUNTS
79
84
if acc["is_valid"]
80
and (
81
acc["error_count"] < max_error_count
82
or now - acc["last_used"] > error_cooldown
83
)
85
and (
86
acc["error_count"] < max_error_count
87
or now - acc["last_used"] > error_cooldown
88
)
84
89
]
85
90
86
91
if not valid_accounts:
@@ -89,8 +94,8 @@ async def get_best_yupp_account() -> Optional[YUPP_ACCOUNT]:
89
94
# Reset error count for accounts in cooldown
90
95
for acc in valid_accounts:
91
96
if (
92
acc["error_count"] >= max_error_count
93
and now - acc["last_used"] > error_cooldown
97
acc["error_count"] >= max_error_count
98
and now - acc["last_used"] > error_cooldown
94
99
):
95
100
acc["error_count"] = 0
96
101
@@ -100,6 +105,7 @@ async def get_best_yupp_account() -> Optional[YUPP_ACCOUNT]:
100
105
account["last_used"] = now
101
106
return account
102
107
108
103
109
async def claim_yupp_reward(session: aiohttp.ClientSession, account: YUPP_ACCOUNT, reward_id: str):
104
110
"""Claim Yupp reward asynchronously"""
105
111
try:
@@ -109,7 +115,7 @@ async def claim_yupp_reward(session: aiohttp.ClientSession, account: YUPP_ACCOUN
109
115
headers = {
110
116
"Content-Type": "application/json",
111
117
"Cookie": f"__Secure-yupp.session-token={account['token']}",
112
"User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/137.0.0.0 Safari/537.36 Edg/137.0.0.0",
118
"User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/137.0.0.0 Safari/537.36 Edg/137.0.0.0",
113
119
114
120
}
115
121
async with session.post(url, json=payload, headers=headers) as response:
@@ -122,6 +128,7 @@ async def claim_yupp_reward(session: aiohttp.ClientSession, account: YUPP_ACCOUN
122
128
log_debug(f"Failed to claim reward {reward_id}. Error: {e}")
123
129
return None
124
130
131
125
132
async def make_chat_private(session: aiohttp.ClientSession, account: YUPP_ACCOUNT, chat_id: str) -> bool:
126
133
"""Set a Yupp chat's sharing status to PRIVATE"""
127
134
try:
@@ -138,7 +145,7 @@ async def make_chat_private(session: aiohttp.ClientSession, account: YUPP_ACCOUN
138
145
headers = {
139
146
"Content-Type": "application/json",
140
147
"Cookie": f"__Secure-yupp.session-token={account['token']}",
141
"User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/137.0.0.0 Safari/537.36 Edg/137.0.0.0",
148
"User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/137.0.0.0 Safari/537.36 Edg/137.0.0.0",
142
149
143
150
}
144
151
@@ -146,8 +153,8 @@ async def make_chat_private(session: aiohttp.ClientSession, account: YUPP_ACCOUN
146
153
response.raise_for_status()
147
154
data = await response.json()
148
155
if (
149
isinstance(data, list) and len(data) > 0
150
and "json" in data[0].get("result", {}).get("data", {})
156
isinstance(data, list) and len(data) > 0
157
and "json" in data[0].get("result", {}).get("data", {})
151
158
):
152
159
log_debug(f"Chat {chat_id} is now PRIVATE ✅")
153
160
return True
@@ -159,6 +166,7 @@ async def make_chat_private(session: aiohttp.ClientSession, account: YUPP_ACCOUN
159
166
log_debug(f"Failed to make chat {chat_id} private: {e}")
160
167
return False
161
168
169
162
170
def log_debug(message: str):
163
171
"""Debug logging"""
164
172
if os.getenv("DEBUG_MODE", "false").lower() == "true":
@@ -166,11 +174,12 @@ def log_debug(message: str):
166
174
else:
167
175
log(f"[Yupp] {message}")
168
176
177
169
178
def format_messages_for_yupp(messages: Messages) -> str:
170
179
"""Format multi-turn conversation for Yupp single-turn format"""
171
180
if not messages:
172
181
return ""
173
182
174
183
if len(messages) == 1 and isinstance(messages[0].get("content"), str):
175
184
return messages[0].get("content", "").strip()
176
185
@@ -202,6 +211,7 @@ def format_messages_for_yupp(messages: Messages) -> str:
202
211
203
212
return result
204
213
214
205
215
class Yupp(AsyncGeneratorProvider, ProviderModelMixin):
206
216
"""
207
217
Yupp.ai Provider for g4f
@@ -214,7 +224,7 @@ class Yupp(AsyncGeneratorProvider, ProviderModelMixin):
214
224
active_by_default = True
215
225
supports_stream = True
216
226
image_cache = True
217
227
218
228
@classmethod
219
229
def get_models(cls, api_key: str = None, **kwargs) -> List[str]:
220
230
if not cls.models:
@@ -230,11 +240,12 @@ class Yupp(AsyncGeneratorProvider, ProviderModelMixin):
230
240
cls.models_tags = {model.get("name"): manager.processor.generate_tags(model) for model in models}
231
241
cls.models = [model.get("name") for model in models]
232
242
cls.image_models = [model.get("name") for model in models if model.get("isImageGeneration")]
233
cls.vision_models = [model.get("name") for model in models if "image/*" in model.get("supportedAttachmentMimeTypes", [])]
243
cls.vision_models = [model.get("name") for model in models if
244
"image/*" in model.get("supportedAttachmentMimeTypes", [])]
234
245
return cls.models
235
246
236
247
@classmethod
237
async def prepare_files(cls, media, session:aiohttp.ClientSession, account:YUPP_ACCOUNT)->list:
248
async def prepare_files(cls, media, session: aiohttp.ClientSession, account: YUPP_ACCOUNT) -> list:
238
249
files = []
239
250
if not media:
240
251
return files
@@ -291,11 +302,11 @@ class Yupp(AsyncGeneratorProvider, ProviderModelMixin):
291
302
292
303
@classmethod
293
304
async def create_async_generator(
294
cls,
295
model: str,
296
messages: Messages,
297
proxy: str = None,
298
**kwargs,
305
cls,
306
model: str,
307
messages: Messages,
308
proxy: str = None,
309
**kwargs,
299
310
) -> AsyncResult:
300
311
"""
301
312
Create async completion using Yupp.ai API with account rotation
@@ -314,15 +325,16 @@ class Yupp(AsyncGeneratorProvider, ProviderModelMixin):
314
325
conversation = kwargs.get("conversation")
315
326
url_uuid = conversation.url_uuid if conversation else None
316
327
is_new_conversation = url_uuid is None
317
328
318
329
prompt = kwargs.get("prompt")
319
330
if prompt is None:
320
331
if is_new_conversation:
321
332
prompt = format_messages_for_yupp(messages)
322
333
else:
323
334
prompt = get_last_user_message(messages, prompt)
324
325
log_debug(f"Use url_uuid: {url_uuid}, Formatted prompt length: {len(prompt)}, Is new conversation: {is_new_conversation}")
335
336
log_debug(
337
f"Use url_uuid: {url_uuid}, Formatted prompt length: {len(prompt)}, Is new conversation: {is_new_conversation}")
326
338
327
339
# Try all accounts with rotation
328
340
max_attempts = len(YUPP_ACCOUNTS)
@@ -332,10 +344,9 @@ class Yupp(AsyncGeneratorProvider, ProviderModelMixin):
332
344
raise ProviderException("No valid Yupp accounts available")
333
345
334
346
try:
335
async with aiohttp.ClientSession() as session:
347
async with StreamSession() as session:
336
348
turn_id = str(uuid.uuid4())
337
349
338
339
350
# Handle media attachments
340
351
media = kwargs.get("media")
341
352
if media:
@@ -390,16 +401,23 @@ class Yupp(AsyncGeneratorProvider, ProviderModelMixin):
390
401
391
402
log_debug(f"Sending request to: {url}")
392
403
log_debug(f"Payload structure: {type(payload)}, length: {len(str(payload))}")
393
404
_timeout = kwargs.get("timeout")
405
if isinstance(_timeout, aiohttp.ClientTimeout):
406
timeout = _timeout
407
else:
408
total = float(_timeout) if isinstance(_timeout, (int, float)) else 5 * 60
409
timeout = aiohttp.ClientTimeout(total=total)
394
410
# Send request
395
async with session.post(url, json=payload, headers=headers, proxy=proxy) as response:
411
async with session.post(url, json=payload, headers=headers, proxy=proxy,
412
timeout=timeout) as response:
396
413
response.raise_for_status()
397
414
398
415
# Make chat private in background
399
416
asyncio.create_task(make_chat_private(session, account, url_uuid))
400
417
401
418
# Process stream
402
async for chunk in cls._process_stream_response(response.content, account, session, prompt, model):
419
async for chunk in cls._process_stream_response(response.content, account, session, prompt,
420
model):
403
421
yield chunk
404
422
405
423
return
@@ -417,7 +435,7 @@ class Yupp(AsyncGeneratorProvider, ProviderModelMixin):
417
435
else:
418
436
account["error_count"] += 1
419
437
continue
420
except ClientResponseError as e:
438
except aiohttp.ClientResponseError as e:
421
439
log_debug(f"Account ...{account['token'][-4:]} failed: {str(e)}")
422
440
# No Available Yupp credits
423
441
if e.status == 500 and 'Internal Server Error' in e.message:
@@ -439,12 +457,12 @@ class Yupp(AsyncGeneratorProvider, ProviderModelMixin):
439
457
440
458
@classmethod
441
459
async def _process_stream_response(
442
cls,
443
response_content,
444
account: YUPP_ACCOUNT,
445
session: aiohttp.ClientSession,
446
prompt: str,
447
model_id: str
460
cls,
461
response_content,
462
account: YUPP_ACCOUNT,
463
session: aiohttp.ClientSession,
464
prompt: str,
465
model_id: str
448
466
) -> AsyncResult:
449
467
"""Process Yupp stream response asynchronously"""
450
468
@@ -461,15 +479,33 @@ class Yupp(AsyncGeneratorProvider, ProviderModelMixin):
461
479
"target": [],
462
480
"variant": [],
463
481
"quick": [],
464
"thinking": [] ,
482
"thinking": [],
465
483
"extra": []
466
484
}
467
485
# Holds leftStream / rightStream definitions to determine target/variant
468
486
select_stream = [None, None]
469
487
# State for capturing a multi-line <think> + <yapp> block (fa-style)
488
capturing_ref_id: Optional[str] = None
489
capturing_lines: List[bytes] = []
490
491
# Storage for special referenced blocks like $fa
492
think_blocks: Dict[str, str] = {}
493
image_blocks: Dict[str, str] = {}
494
470
495
def extract_ref_id(ref):
471
496
"""Extract ID from reference string, e.g., from '$@123' extract '123'"""
472
497
return ref[2:] if ref and isinstance(ref, str) and ref.startswith("$@") else None
498
499
def extract_ref_name(ref: str) -> Optional[str]:
500
"""Extract simple ref name from '$fa' → 'fa'"""
501
if not isinstance(ref, str):
502
return None
503
if ref.startswith("$@"):
504
return ref[2:]
505
if ref.startswith("$") and len(ref) > 1:
506
return ref[1:]
507
return None
508
473
509
def is_valid_content(content: str) -> bool:
474
510
"""Check if content is valid"""
475
511
if not content or content in [None, "", "$undefined"]:
@@ -515,7 +551,27 @@ class Yupp(AsyncGeneratorProvider, ProviderModelMixin):
515
551
if for_target:
516
552
normal_content += content
517
553
yield content
518
554
555
def finalize_capture_block(ref_id: str, lines: List[bytes]):
556
"""Parse captured <think> + <yapp> block for a given ref ID."""
557
text = b"".join(lines).decode("utf-8", errors="ignore")
558
559
# Extract <think>...</think>
560
think_start = text.find("<think>")
561
think_end = text.find("</think>")
562
if think_start != -1 and think_end != -1 and think_end > think_start:
563
inner = text[think_start + len("<think>"):think_end].strip()
564
if inner:
565
think_blocks[ref_id] = inner
566
567
# Extract <yapp class="image-gen">...</yapp>
568
yapp_start = text.find('<yapp class="image-gen">')
569
if yapp_start != -1:
570
yapp_end = text.find("</yapp>", yapp_start)
571
if yapp_end != -1:
572
yapp_block = text[yapp_start:yapp_end + len("</yapp>")]
573
image_blocks[ref_id] = yapp_block
574
519
575
try:
520
576
line_count = 0
521
577
quick_response_id = None
@@ -531,17 +587,55 @@ class Yupp(AsyncGeneratorProvider, ProviderModelMixin):
531
587
right_message_id = None
532
588
nudge_new_chat_id = None
533
589
nudge_new_chat = False
534
535
590
async for line in response_content:
536
591
line_count += 1
537
592
# If we are currently capturing a think/image block for some ref ID
593
if capturing_ref_id is not None:
594
capturing_lines.append(line)
595
596
# Check if this line closes the <yapp> block; after that, block is complete
597
if b"</yapp>" in line: # or b':{"curr"' in line:
598
# We may have trailing "2:{...}" after </yapp> on the same line
599
# Get id using re
600
idx = line.find(b"</yapp>")
601
suffix = line[idx + len(b"</yapp>"):]
602
603
# Finalize captured block for this ref ID
604
finalize_capture_block(capturing_ref_id, capturing_lines)
605
capturing_ref_id = None
606
capturing_lines = []
607
608
# If there is trailing content (e.g. '2:{"curr":"$fa"...}')
609
if suffix.strip():
610
# Process suffix as a new "line" in the same iteration
611
line = suffix
612
else:
613
# Nothing more on this line
614
continue
615
else:
616
# Still inside captured block; skip normal processing
617
continue
618
619
# Detect start of a <think> block assigned to a ref like 'fa:...<think>'
620
if b"<think>" in line:
621
m = line_pattern.match(line)
622
if m:
623
capturing_ref_id = m.group(1).decode()
624
capturing_lines = [line]
625
# Skip normal parsing; the rest of the block will be captured until </yapp>
626
continue
627
538
628
match = line_pattern.match(line)
539
629
if not match:
540
630
continue
541
631
542
632
chunk_id, chunk_data = match.groups()
543
633
chunk_id = chunk_id.decode()
544
634
635
if nudge_new_chat_id and chunk_id == nudge_new_chat_id:
636
nudge_new_chat = chunk_data.decode()
637
continue
638
545
639
try:
546
640
data = json.loads(chunk_data) if chunk_data != b"{}" else {}
547
641
except json.JSONDecodeError:
@@ -550,7 +644,7 @@ class Yupp(AsyncGeneratorProvider, ProviderModelMixin):
550
644
if chunk_id == reward_id and isinstance(data, dict) and "unclaimedRewardInfo" in data:
551
645
reward_info = data
552
646
log_debug(f"Found reward info")
553
647
554
648
# Process initial setup
555
649
elif chunk_id == "1":
556
650
yield PlainTextResponse(line.decode(errors="ignore"))
@@ -558,7 +652,8 @@ class Yupp(AsyncGeneratorProvider, ProviderModelMixin):
558
652
left_stream = data.get("leftStream", {})
559
653
right_stream = data.get("rightStream", {})
560
654
if data.get("quickResponse", {}) != "$undefined":
561
quick_response_id = extract_ref_id(data.get("quickResponse", {}).get("stream", {}).get("next"))
655
quick_response_id = extract_ref_id(
656
data.get("quickResponse", {}).get("stream", {}).get("next"))
562
657
563
658
if data.get("turnId", {}) != "$undefined":
564
659
turn_id = extract_ref_id(data.get("turnId", {}).get("next"))
@@ -592,7 +687,7 @@ class Yupp(AsyncGeneratorProvider, ProviderModelMixin):
592
687
provider_info["variantUrl"] = selection.get("externalUrl")
593
688
log_debug(f"Found variant stream ID: {variant_stream_id}")
594
689
yield ProviderInfo.from_dict(provider_info)
595
690
596
691
# Process target stream content
597
692
elif target_stream_id and chunk_id == target_stream_id:
598
693
yield PlainTextResponse(line.decode(errors="ignore"))
@@ -600,15 +695,41 @@ class Yupp(AsyncGeneratorProvider, ProviderModelMixin):
600
695
target_stream_id = extract_ref_id(data.get("next"))
601
696
content = data.get("curr", "")
602
697
if content:
603
async for chunk in process_content_chunk(
604
content,
605
chunk_id,
606
line_count,
607
for_target=True
608
):
609
stream["target"].append(chunk)
610
is_started = True
611
yield chunk
698
# Handle special "$fa" / "$<id>" reference
699
ref_name = extract_ref_name(content)
700
if ref_name and (ref_name in think_blocks or ref_name in image_blocks):
701
# Thinking block
702
if ref_name in think_blocks:
703
t_text = think_blocks[ref_name]
704
if t_text:
705
reasoning = Reasoning(t_text)
706
# thinking_content += t_text
707
stream["thinking"].append(reasoning)
708
# yield reasoning
709
710
# Image-gen block
711
if ref_name in image_blocks:
712
img_block_text = image_blocks[ref_name]
713
async for chunk in process_content_chunk(
714
img_block_text,
715
ref_name,
716
line_count,
717
for_target=True
718
):
719
stream["target"].append(chunk)
720
is_started = True
721
yield chunk
722
else:
723
# Normal textual chunk
724
async for chunk in process_content_chunk(
725
content,
726
chunk_id,
727
line_count,
728
for_target=True
729
):
730
stream["target"].append(chunk)
731
is_started = True
732
yield chunk
612
733
# Variant stream (comparison)
613
734
elif variant_stream_id and chunk_id == variant_stream_id:
614
735
yield PlainTextResponse("[Variant] " + line.decode(errors="ignore"))
@@ -651,8 +772,6 @@ class Yupp(AsyncGeneratorProvider, ProviderModelMixin):
651
772
...
652
773
elif chunk_id == left_message_id:
653
774
...
654
elif chunk_id == nudge_new_chat_id:
655
nudge_new_chat = data
656
775
# Miscellaneous extra content
657
776
elif isinstance(data, dict) and "curr" in data:
658
777
content = data.get("curr", "")
@@ -664,24 +783,23 @@ class Yupp(AsyncGeneratorProvider, ProviderModelMixin):
664
783
for_target=False
665
784
):
666
785
stream["extra"].append(chunk)
667
if isinstance(chunk,str) and "<streaming stopped unexpectedly" in chunk:
786
if isinstance(chunk, str) and "<streaming stopped unexpectedly" in chunk:
668
787
yield FinishReason(chunk)
669
788
670
789
yield PlainTextResponse("[Extra] " + line.decode(errors="ignore"))
671
790
672
791
if variant_image is not None:
673
792
yield variant_image
674
793
elif variant_text:
675
794
yield PreviewResponse(variant_text)
676
795
yield JsonResponse(**stream)
677
796
log_debug(f"Finished processing {line_count} lines")
678
679
797
except:
680
798
raise
681
799
682
800
finally:
683
801
# Claim reward in background
684
802
if reward_info and "unclaimedRewardInfo" in reward_info:
685
803
reward_id = reward_info["unclaimedRewardInfo"].get("rewardId")
686
804
if reward_id:
687
await claim_yupp_reward(session, account, reward_id)
805
await claim_yupp_reward(session, account, reward_id)
@@ -642,7 +642,7 @@ class LMArena(AsyncGeneratorProvider, ProviderModelMixin, AuthFileMixin):
642
642
643
643
if not cls._models_loaded:
644
644
cls.get_models()
645
is_image_model = model in image_models
645
is_image_model = model in cls.image_models
646
646
if not model:
647
647
model = cls.default_model
648
648
if model in cls.model_aliases:
@@ -1,18 +1,20 @@
1
1
from __future__ import annotations
2
2
3
import os
4
import re
5
3
import asyncio
6
import uuid
7
import json
8
4
import base64
9
import time
5
import hashlib
6
import json
7
import os
10
8
import random
11
from typing import AsyncIterator, Iterator, Optional, Generator, Dict, Union, List, Any
9
import re
10
import time
11
import uuid
12
12
from copy import copy
13
from typing import AsyncIterator, Iterator, Optional, Generator, Dict, Union, List, Any
13
14
14
15
try:
15
16
import nodriver
17
16
18
has_nodriver = True
17
19
except ImportError:
18
20
has_nodriver = False
@@ -22,15 +24,17 @@ from ...typing import AsyncResult, Messages, Cookies, MediaListType
22
24
from ...requests.raise_for_status import raise_for_status
23
25
from ...requests import StreamSession
24
26
from ...requests import get_nodriver_session
25
from ...image import ImageRequest, to_image, to_bytes, is_accepted_format, detect_file_type
27
from ...image import ImageRequest, to_image, to_bytes, detect_file_type
26
28
from ...errors import MissingAuthError, NoValidHarFileError, ModelNotFoundError
27
from ...providers.response import JsonConversation, FinishReason, SynthesizeData, AuthResult, ImageResponse, ImagePreview, ResponseType, JsonRequest, format_link
29
from ...providers.response import JsonConversation, FinishReason, SynthesizeData, AuthResult, ImageResponse, \
30
ImagePreview, ResponseType, JsonRequest, format_link
28
31
from ...providers.response import TitleGeneration, RequestLogin, Reasoning
29
32
from ...tools.media import merge_media
30
33
from ..helper import format_cookies, format_media_prompt, to_string
31
34
from ..openai.models import default_model, default_image_model, models, image_models, text_models, model_aliases
32
35
from ..openai.har_file import get_request_config
33
from ..openai.har_file import RequestConfig, arkReq, arkose_url, start_url, conversation_url, backend_url, prepare_url, backend_anon_url
36
from ..openai.har_file import RequestConfig, arkReq, arkose_url, start_url, conversation_url, backend_url, prepare_url, \
37
backend_anon_url
34
38
from ..openai.proofofwork import generate_proof_token
35
39
from ..openai.new import get_requirements_token, get_config
36
40
from ... import debug
@@ -87,6 +91,9 @@ UPLOAD_HEADERS = {
87
91
"user-agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/131.0.0.0 Safari/537.36"
88
92
}
89
93
94
ImagesCache: Dict[str, dict] = {}
95
96
90
97
class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
91
98
"""A class for creating and managing conversations with OpenAI chat service"""
92
99
@@ -95,6 +102,7 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
95
102
working = True
96
103
active_by_default = True
97
104
use_nodriver = True
105
image_cache = True
98
106
supports_gpt_4 = True
99
107
supports_message_history = True
100
108
supports_system_message = True
@@ -127,11 +135,11 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
127
135
128
136
@classmethod
129
137
async def upload_files(
130
cls,
131
session: StreamSession,
132
auth_result: AuthResult,
133
media: MediaListType,
134
) -> list[ImageRequest]:
138
cls,
139
session: StreamSession,
140
auth_result: AuthResult,
141
media: MediaListType,
142
) -> List[ImageRequest]:
135
143
"""
136
144
Upload an image to the service and get the download URL
137
145
@@ -143,11 +151,20 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
143
151
Returns:
144
152
An ImageRequest object that contains the download URL, file name, and other data
145
153
"""
146
async def upload_file(file, image_name=None):
154
155
async def upload_file(file, image_name=None) -> ImageRequest:
147
156
debug.log(f"Uploading file: {image_name}")
148
157
file_data = {}
149
158
150
159
data_bytes = to_bytes(file)
160
# Check Cache
161
hasher = hashlib.md5()
162
hasher.update(data_bytes)
163
image_hash = hasher.hexdigest()
164
cache_file = ImagesCache.get(image_hash)
165
if cls.image_cache and file:
166
debug.log("Using cached image")
167
return ImageRequest(cache_file)
151
168
extension, mime_type = detect_file_type(data_bytes)
152
169
if "image" in mime_type:
153
170
# Convert the image to a PIL Image object
@@ -181,30 +198,31 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
181
198
# Put the image bytes to the upload URL and check the status
182
199
await asyncio.sleep(1)
183
200
async with session.put(
184
file_data["upload_url"],
185
data=data_bytes,
186
headers={
187
**UPLOAD_HEADERS,
188
"Content-Type": file_data["mime_type"],
189
"x-ms-blob-type": "BlockBlob",
190
"x-ms-version": "2020-04-08",
191
"Origin": "https://chatgpt.com",
192
}
201
file_data["upload_url"],
202
data=data_bytes,
203
headers={
204
**UPLOAD_HEADERS,
205
"Content-Type": file_data["mime_type"],
206
"x-ms-blob-type": "BlockBlob",
207
"x-ms-version": "2020-04-08",
208
"Origin": "https://chatgpt.com",
209
}
193
210
) as response:
194
211
await raise_for_status(response)
195
212
# Post the file ID to the service and get the download URL
196
213
async with session.post(
197
f"{cls.url}/backend-api/files/{file_data['file_id']}/uploaded",
198
json={},
199
headers=auth_result.headers
214
f"{cls.url}/backend-api/files/{file_data['file_id']}/uploaded",
215
json={},
216
headers=auth_result.headers
200
217
) as response:
201
218
cls._update_request_args(auth_result, session)
202
219
await raise_for_status(response, "Get download url failed")
203
220
uploaded_data = await response.json()
204
221
file_data["download_url"] = uploaded_data["download_url"]
222
ImagesCache[image_hash] = file_data.copy()
205
223
return ImageRequest(file_data)
206
224
207
medias = []
225
medias: List["ImageRequest"] = []
208
226
for item in media:
209
227
item = item if isinstance(item, tuple) else (item,)
210
228
__uploaded_media = await upload_file(*item)
@@ -242,7 +260,8 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
242
260
"id": str(uuid.uuid4()),
243
261
"author": {"role": message["role"]},
244
262
"content": {"content_type": "text", "parts": [to_string(message["content"])]},
245
"metadata": {"serialization_metadata": {"custom_symbol_offsets": []}, **({"system_hints": system_hints} if system_hints else {})},
263
"metadata": {"serialization_metadata": {"custom_symbol_offsets": []},
264
**({"system_hints": system_hints} if system_hints else {})},
246
265
"create_time": time.time(),
247
266
} for message in messages]
248
267
# Check if there is an image response
@@ -256,11 +275,11 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
256
275
"size_bytes": image_request.get("file_size"),
257
276
"width": image_request.get("width"),
258
277
}
259
for image_request in image_requests
278
for image_request in image_requests
260
279
# Add For Images Only
261
280
if image_request.get("use_case") == "multimodal"
262
281
],
263
messages[-1]["content"]["parts"][0]]
282
messages[-1]["content"]["parts"][0]]
264
283
}
265
284
# Add the metadata object with the attachments
266
285
messages[-1]["metadata"] = {
@@ -278,12 +297,14 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
278
297
else {}
279
298
),
280
299
}
281
for image_request in image_requests]
300
for image_request in image_requests]
282
301
}
283
302
return messages
284
303
285
304
@classmethod
286
async def get_generated_image(cls, session: StreamSession, auth_result: AuthResult, element: Union[dict, str], prompt: str = None, conversation_id: str = None) -> ImagePreview|ImageResponse|None:
305
async def get_generated_image(cls, session: StreamSession, auth_result: AuthResult, element: Union[dict, str],
306
prompt: str = None, conversation_id: str = None,
307
status: Optional[str] = None) -> ImagePreview | ImageResponse | None:
287
308
download_urls = []
288
309
is_sediment = False
289
310
if prompt is None:
@@ -292,7 +313,7 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
292
313
except KeyError:
293
314
pass
294
315
if "asset_pointer" in element:
295
element = element["asset_pointer"]
316
element = element["asset_pointer"]
296
317
if isinstance(element, str) and element.startswith("file-service://"):
297
318
element = element.split("file-service://", 1)[-1]
298
319
elif isinstance(element, str) and element.startswith("sediment://"):
@@ -303,7 +324,7 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
303
324
if is_sediment:
304
325
url = f"{cls.url}/backend-api/conversation/{conversation_id}/attachment/{element}/download"
305
326
else:
306
url =f"{cls.url}/backend-api/files/{element}/download"
327
url = f"{cls.url}/backend-api/files/{element}/download"
307
328
try:
308
329
async with session.get(url, headers=auth_result.headers) as response:
309
330
cls._update_request_args(auth_result, session)
@@ -319,27 +340,31 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
319
340
debug.error("OpenaiChat: Download image failed")
320
341
debug.error(e)
321
342
if download_urls:
322
return ImagePreview(download_urls, prompt, {"headers": auth_result.headers}) if is_sediment else ImageResponse(download_urls, prompt, {"headers": auth_result.headers})
343
# status = None, finished_successfully
344
if is_sediment and status is None:
345
return ImagePreview(download_urls, prompt, {"status": status, "headers": auth_result.headers})
346
else:
347
return ImageResponse(download_urls, prompt, {"status": status, "headers": auth_result.headers})
323
348
324
349
@classmethod
325
350
async def create_authed(
326
cls,
327
model: str,
328
messages: Messages,
329
auth_result: AuthResult,
330
proxy: str = None,
331
timeout: int = 360,
332
auto_continue: bool = False,
333
action: Optional[str] = None,
334
conversation: Conversation = None,
335
media: MediaListType = None,
336
return_conversation: bool = True,
337
web_search: bool = False,
338
prompt: str = None,
339
conversation_mode: Optional[dict] = None,
340
temporary: Optional[bool] = None,
341
conversation_id: Optional[str] = None,
342
**kwargs
351
cls,
352
model: str,
353
messages: Messages,
354
auth_result: AuthResult,
355
proxy: str = None,
356
timeout: int = 360,
357
auto_continue: bool = False,
358
action: Optional[str] = None,
359
conversation: Conversation = None,
360
media: MediaListType = None,
361
return_conversation: bool = True,
362
web_search: bool = False,
363
prompt: str = None,
364
conversation_mode: Optional[dict] = None,
365
temporary: Optional[bool] = None,
366
conversation_id: Optional[str] = None,
367
**kwargs
343
368
) -> AsyncResult:
344
369
"""
345
370
Create an asynchronous generator for the conversation.
@@ -367,12 +392,12 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
367
392
if action is None:
368
393
action = "next"
369
394
async with StreamSession(
370
proxy=proxy,
371
impersonate="chrome",
372
timeout=timeout
395
proxy=proxy,
396
impersonate="chrome",
397
timeout=timeout
373
398
) as session:
374
399
image_requests = None
375
media = merge_media(media, messages)
400
media = merge_media(media, messages)
376
401
if not cls.needs_auth and not media:
377
402
if cls._headers is None:
378
403
cls._create_request_args(cls._cookies)
@@ -436,18 +461,19 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
436
461
if temporary:
437
462
data["history_and_training_disabled"] = True
438
463
async with session.post(
439
prepare_url,
440
json=data,
441
headers=cls._headers
464
prepare_url,
465
json=data,
466
headers=cls._headers
442
467
) as response:
443
468
await raise_for_status(response)
444
469
conduit_token = (await response.json())["conduit_token"]
445
470
async with session.post(
446
f"{cls.url}/backend-anon/sentinel/chat-requirements"
447
if cls._api_key is None else
448
f"{cls.url}/backend-api/sentinel/chat-requirements",
449
json={"p": None if not getattr(auth_result, "proof_token", None) else get_requirements_token(getattr(auth_result, "proof_token", None))},
450
headers=cls._headers
471
f"{cls.url}/backend-anon/sentinel/chat-requirements"
472
if cls._api_key is None else
473
f"{cls.url}/backend-api/sentinel/chat-requirements",
474
json={"p": None if not getattr(auth_result, "proof_token", None) else get_requirements_token(
475
getattr(auth_result, "proof_token", None))},
476
headers=cls._headers
451
477
) as response:
452
478
if response.status in (401, 403):
453
479
raise MissingAuthError(f"Response status: {response.status}")
@@ -456,10 +482,10 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
456
482
await raise_for_status(response)
457
483
chat_requirements = await response.json()
458
484
need_turnstile = chat_requirements.get("turnstile", {}).get("required", False)
459
need_arkose = chat_requirements.get("arkose", {}).get("required", False)
460
chat_token = chat_requirements.get("token")
485
need_arkose = chat_requirements.get("arkose", {}).get("required", False)
486
chat_token = chat_requirements.get("token")
461
487
462
# if need_arkose and cls.request_config.arkose_token is None:
488
# if need_arkose and cls.request_config.arkose_token is None:
463
489
# await get_request_config(proxy)
464
490
# cls._create_request_args(auth_result.cookies, auth_result.headers)
465
491
# cls._set_api_key(auth_result.access_token)
@@ -476,23 +502,25 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
476
502
proof_token=proof_token
477
503
)
478
504
# [debug.log(text) for text in (
479
#f"Arkose: {'False' if not need_arkose else auth_result.arkose_token[:12]+'...'}",
480
#f"Proofofwork: {'False' if proofofwork is None else proofofwork[:12]+'...'}",
481
#f"AccessToken: {'False' if cls._api_key is None else cls._api_key[:12]+'...'}",
505
# f"Arkose: {'False' if not need_arkose else auth_result.arkose_token[:12]+'...'}",
506
# f"Proofofwork: {'False' if proofofwork is None else proofofwork[:12]+'...'}",
507
# f"AccessToken: {'False' if cls._api_key is None else cls._api_key[:12]+'...'}",
482
508
# )]
483
509
data = {
484
510
"action": "next",
485
511
"parent_message_id": conversation.message_id,
486
512
"model": model,
487
"timezone_offset_min":-120,
488
"timezone":"Europe/Berlin",
489
"conversation_mode":{"kind":"primary_assistant"},
490
"enable_message_followups":True,
513
"timezone_offset_min": -120,
514
"timezone": "Europe/Berlin",
515
"conversation_mode": {"kind": "primary_assistant"},
516
"enable_message_followups": True,
491
517
"system_hints": ["search"] if web_search else None,
492
"supports_buffering":True,
493
"supported_encodings":["v1"],
494
"client_contextual_info":{"is_dark_mode":False,"time_since_loaded":random.randint(20, 500),"page_height":578,"page_width":1850,"pixel_ratio":1,"screen_height":1080,"screen_width":1920},
495
"paragen_cot_summary_display_override":"allow"
518
"supports_buffering": True,
519
"supported_encodings": ["v1"],
520
"client_contextual_info": {"is_dark_mode": False, "time_since_loaded": random.randint(20, 500),
521
"page_height": 578, "page_width": 1850, "pixel_ratio": 1,
522
"screen_height": 1080, "screen_width": 1920},
523
"paragen_cot_summary_display_override": "allow"
496
524
}
497
525
if temporary:
498
526
data["history_and_training_disabled"] = True
@@ -512,7 +540,8 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
512
540
new_messages = []
513
541
else:
514
542
new_messages.append(message)
515
data["messages"] = cls.create_messages(new_messages, image_requests, ["search"] if web_search else None)
543
data["messages"] = cls.create_messages(new_messages, image_requests,
544
["search"] if web_search else None)
516
545
yield JsonRequest.from_dict(data)
517
546
headers = {
518
547
**cls._headers,
@@ -521,18 +550,18 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
521
550
"openai-sentinel-chat-requirements-token": chat_token,
522
551
**({} if conduit_token is None else {"x-conduit-token": conduit_token})
523
552
}
524
#if cls.request_config.arkose_token:
553
# if cls.request_config.arkose_token:
525
554
# headers["openai-sentinel-arkose-token"] = cls.request_config.arkose_token
526
555
if proofofwork is not None:
527
556
headers["openai-sentinel-proof-token"] = proofofwork
528
557
if need_turnstile and getattr(auth_result, "turnstile_token", None) is not None:
529
558
headers['openai-sentinel-turnstile-token'] = auth_result.turnstile_token
530
559
async with session.post(
531
backend_anon_url
532
if cls._api_key is None else
533
backend_url,
534
json=data,
535
headers=headers
560
backend_anon_url
561
if cls._api_key is None else
562
backend_url,
563
json=data,
564
headers=headers
536
565
) as response:
537
566
cls._update_request_args(auth_result, session)
538
567
if response.status in (401, 403, 429, 500):
@@ -548,10 +577,12 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
548
577
if match.group(0) in matches:
549
578
continue
550
579
matches.append(match.group(0))
551
generated_image = await cls.get_generated_image(session, auth_result, match.group(0), prompt)
580
generated_image = await cls.get_generated_image(session, auth_result, match.group(0),
581
prompt)
552
582
if generated_image is not None:
553
583
yield generated_image
554
async for chunk in cls.iter_messages_line(session, auth_result, line, conversation, sources, references):
584
async for chunk in cls.iter_messages_line(session, auth_result, line, conversation, sources,
585
references):
555
586
if isinstance(chunk, str):
556
587
chunk = chunk.replace("\ue203", "").replace("\ue204", "").replace("\ue206", "")
557
588
buffer += chunk
@@ -561,9 +592,9 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
561
592
def citation_replacer(match: re.Match[str]):
562
593
ref_type = match.group(1)
563
594
ref_index = int(match.group(2))
564
if ((ref_type == "image" and is_image_embedding) or
565
is_video_embedding or
566
ref_type == "forecast"):
595
if ((ref_type == "image" and is_image_embedding) or
596
is_video_embedding or
597
ref_type == "forecast"):
567
598
568
599
reference = references.get_reference({
569
600
"ref_index": ref_index,
@@ -571,7 +602,7 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
571
602
})
572
603
if not reference:
573
604
return ""
574
605
575
606
if ref_type == "forecast":
576
607
if reference.get("alt"):
577
608
return reference.get("alt")
@@ -580,11 +611,13 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
580
611
581
612
if is_image_embedding and reference.get("content_url", ""):
582
613
return f"})"
583
614
584
615
if is_video_embedding:
585
if reference.get("url", "") and reference.get("thumbnail_url", ""):
616
if reference.get("url", "") and reference.get("thumbnail_url",
617
""):
586
618
return f"[]({reference['url']})"
587
video_match = re.match(r"video\n(.*?)\nturn[0-9]+", match.group(0))
619
video_match = re.match(r"video\n(.*?)\nturn[0-9]+",
620
match.group(0))
588
621
if video_match:
589
622
return video_match.group(1)
590
623
return ""
@@ -595,9 +628,9 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
595
628
})
596
629
if source_index is not None and len(sources.list) > source_index:
597
630
link = sources.list[source_index]["url"]
598
return f"[[{source_index+1}]]({link})"
631
return f"[[{source_index + 1}]]({link})"
599
632
return f""
600
633
601
634
def products_replacer(match: re.Match[str]):
602
635
try:
603
636
products_data = json.loads(match.group(1))
@@ -612,25 +645,30 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
612
645
return ""
613
646
614
647
sequence_content = match.group(1)
615
sequence_content = sequence_content.replace("\ue200", "").replace("\ue202", "\n").replace("\ue201", "")
648
sequence_content = sequence_content.replace("\ue200", "").replace("\ue202",
649
"\n").replace(
650
"\ue201", "")
616
651
sequence_content = sequence_content.replace("navlist\n", "#### ")
617
652
618
653
# Handle search, news, view and image citations
619
654
is_image_embedding = sequence_content.startswith("i\nturn")
620
655
is_video_embedding = sequence_content.startswith("video\n")
621
656
sequence_content = re.sub(
622
r'(?:cite\nturn[0-9]+|forecast\nturn[0-9]+|video\n.*?\nturn[0-9]+|i?\n?turn[0-9]+)(search|news|view|image|forecast)(\d+)',
623
citation_replacer,
657
r'(?:cite\nturn[0-9]+|forecast\nturn[0-9]+|video\n.*?\nturn[0-9]+|i?\n?turn[0-9]+)(search|news|view|image|forecast)(\d+)',
658
citation_replacer,
624
659
sequence_content
625
660
)
626
sequence_content = re.sub(r'products\n(.*)', products_replacer, sequence_content)
627
sequence_content = re.sub(r'product_entity\n\[".*","(.*)"\]', lambda x: x.group(1), sequence_content)
661
sequence_content = re.sub(r'products\n(.*)', products_replacer,
662
sequence_content)
663
sequence_content = re.sub(r'product_entity\n\[".*","(.*)"\]',
664
lambda x: x.group(1), sequence_content)
628
665
return sequence_content
629
666
630
667
# process only completed sequences and do not touch start of next not completed sequence
631
buffer = re.sub(r'\ue200(.*?)\ue201', sequence_replacer, buffer, flags=re.DOTALL)
632
633
if buffer.find(u"\ue200") != -1: # still have uncompleted sequence
668
buffer = re.sub(r'\ue200(.*?)\ue201', sequence_replacer, buffer,
669
flags=re.DOTALL)
670
671
if buffer.find(u"\ue200") != -1: # still have uncompleted sequence
634
672
continue
635
673
else:
636
674
# do not yield to consume rest part of special sequence
@@ -647,7 +685,8 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
647
685
if sources.list:
648
686
yield sources
649
687
if conversation.generated_images:
650
yield ImageResponse(conversation.generated_images.urls, conversation.prompt, {"headers": auth_result.headers})
688
yield ImageResponse(conversation.generated_images.urls, conversation.prompt,
689
{"headers": auth_result.headers})
651
690
conversation.generated_images = None
652
691
conversation.prompt = None
653
692
if return_conversation:
@@ -667,7 +706,9 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
667
706
yield FinishReason(conversation.finish_reason)
668
707
669
708
@classmethod
670
async def iter_messages_line(cls, session: StreamSession, auth_result: AuthResult, line: bytes, fields: Conversation, sources: OpenAISources, references: ContentReferences) -> AsyncIterator:
709
async def iter_messages_line(cls, session: StreamSession, auth_result: AuthResult, line: bytes,
710
fields: Conversation, sources: OpenAISources,
711
references: ContentReferences) -> AsyncIterator:
671
712
if not line.startswith(b"data: "):
672
713
return
673
714
elif line.startswith(b"data: [DONE]"):
@@ -706,9 +747,14 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
706
747
elif m.get("p") == "/message/metadata/image_gen_title":
707
748
fields.prompt = m.get("v")
708
749
elif m.get("p") == "/message/content/parts/0/asset_pointer":
709
generated_images = fields.generated_images = await cls.get_generated_image(session, auth_result, m.get("v"), fields.prompt, fields.conversation_id)
750
status = next(filter(lambda x: x.get("p") == '/message/status', v), {}).get('v', None)
751
generated_images = fields.generated_images = await cls.get_generated_image(session, auth_result,
752
m.get("v"),
753
fields.prompt,
754
fields.conversation_id,
755
status)
710
756
if generated_images is not None:
711
if buffer:
757
if buffer:
712
758
yield buffer
713
759
yield generated_images
714
760
elif m.get("p") == "/message/metadata/search_result_groups":
@@ -735,41 +781,48 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
735
781
if match and m.get("o") == "append" and isinstance(m.get("v"), dict):
736
782
idx = int(match.group(1))
737
783
references.merge_reference(idx, m.get("v"))
738
elif m.get("p") and re.match(r"^/message/metadata/content_references/\d+/fallback_items$", m.get("p")) and isinstance(m.get("v"), list):
784
elif m.get("p") and re.match(r"^/message/metadata/content_references/\d+/fallback_items$",
785
m.get("p")) and isinstance(m.get("v"), list):
739
786
for link in m.get("v", []) or []:
740
787
sources.add_source(link)
741
elif m.get("p") and re.match(r"^/message/metadata/content_references/\d+/items$", m.get("p")) and isinstance(m.get("v"), list):
788
elif m.get("p") and re.match(r"^/message/metadata/content_references/\d+/items$",
789
m.get("p")) and isinstance(m.get("v"), list):
742
790
for link in m.get("v", []) or []:
743
791
sources.add_source(link)
744
elif m.get("p") and re.match(r"^/message/metadata/content_references/\d+/refs$", m.get("p")) and isinstance(m.get("v"), list):
792
elif m.get("p") and re.match(r"^/message/metadata/content_references/\d+/refs$",
793
m.get("p")) and isinstance(m.get("v"), list):
745
794
match = re.match(r"^/message/metadata/content_references/(\d+)/refs$", m.get("p"))
746
795
if match:
747
796
idx = int(match.group(1))
748
797
references.update_reference(idx, m.get("o"), "refs", m.get("v"))
749
elif m.get("p") and re.match(r"^/message/metadata/content_references/\d+/alt$", m.get("p")) and isinstance(m.get("v"), list):
798
elif m.get("p") and re.match(r"^/message/metadata/content_references/\d+/alt$",
799
m.get("p")) and isinstance(m.get("v"), list):
750
800
match = re.match(r"^/message/metadata/content_references/(\d+)/alt$", m.get("p"))
751
801
if match:
752
802
idx = int(match.group(1))
753
803
references.update_reference(idx, m.get("o"), "alt", m.get("v"))
754
elif m.get("p") and re.match(r"^/message/metadata/content_references/\d+/prompt_text$", m.get("p")) and isinstance(m.get("v"), list):
804
elif m.get("p") and re.match(r"^/message/metadata/content_references/\d+/prompt_text$",
805
m.get("p")) and isinstance(m.get("v"), list):
755
806
match = re.match(r"^/message/metadata/content_references/(\d+)/prompt_text$", m.get("p"))
756
807
if match:
757
808
idx = int(match.group(1))
758
809
references.update_reference(idx, m.get("o"), "prompt_text", m.get("v"))
759
elif m.get("p") and re.match(r"^/message/metadata/content_references/\d+/refs/\d+$", m.get("p")) and isinstance(m.get("v"), dict):
810
elif m.get("p") and re.match(r"^/message/metadata/content_references/\d+/refs/\d+$",
811
m.get("p")) and isinstance(m.get("v"), dict):
760
812
match = re.match(r"^/message/metadata/content_references/(\d+)/refs/(\d+)$", m.get("p"))
761
813
if match:
762
814
reference_idx = int(match.group(1))
763
815
ref_idx = int(match.group(2))
764
816
references.update_reference(reference_idx, m.get("o"), "refs", m.get("v"), ref_idx)
765
elif m.get("p") and re.match(r"^/message/metadata/content_references/\d+/images$", m.get("p")) and isinstance(m.get("v"), list):
817
elif m.get("p") and re.match(r"^/message/metadata/content_references/\d+/images$",
818
m.get("p")) and isinstance(m.get("v"), list):
766
819
match = re.match(r"^/message/metadata/content_references/(\d+)/images$", m.get("p"))
767
820
if match:
768
821
idx = int(match.group(1))
769
822
references.update_reference(idx, m.get("o"), "images", m.get("v"))
770
823
elif m.get("p") == "/message/metadata/finished_text":
771
824
fields.is_thinking = False
772
if buffer:
825
if buffer:
773
826
yield buffer
774
827
yield Reasoning(status=m.get("v"))
775
828
elif m.get("p") == "/message/metadata" and fields.recipient == "all":
@@ -785,10 +838,11 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
785
838
fields.recipient = m.get("recipient", fields.recipient)
786
839
if fields.recipient == "all":
787
840
c = m.get("content", {})
788
if c.get("content_type") == "text" and m.get("author", {}).get("role") == "tool" and "initial_text" in m.get("metadata", {}):
841
if c.get("content_type") == "text" and m.get("author", {}).get(
842
"role") == "tool" and "initial_text" in m.get("metadata", {}):
789
843
fields.is_thinking = True
790
844
yield Reasoning(status=m.get("metadata", {}).get("initial_text"))
791
#if c.get("content_type") == "multimodal_text":
845
# if c.get("content_type") == "multimodal_text":
792
846
# for part in c.get("parts"):
793
847
# if isinstance(part, dict) and part.get("content_type") == "image_asset_pointer":
794
848
# yield await cls.get_generated_image(session, auth_result, part, fields.prompt, fields.conversation_id)
@@ -803,13 +857,13 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
803
857
@classmethod
804
858
async def synthesize(cls, params: dict) -> AsyncIterator[bytes]:
805
859
async with StreamSession(
806
impersonate="chrome",
807
timeout=0
860
impersonate="chrome",
861
timeout=0
808
862
) as session:
809
863
async with session.get(
810
f"{cls.url}/backend-api/synthesize",
811
params=params,
812
headers=cls._headers
864
f"{cls.url}/backend-api/synthesize",
865
params=params,
866
headers=cls._headers
813
867
) as response:
814
868
await raise_for_status(response)
815
869
async for chunk in response.iter_content():
@@ -817,15 +871,15 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
817
871
818
872
@classmethod
819
873
async def login(
820
cls,
821
proxy: str = None,
822
api_key: str = None,
823
proof_token: str = None,
824
cookies: Cookies = None,
825
headers: dict = None,
826
**kwargs
874
cls,
875
proxy: str = None,
876
api_key: str = None,
877
proof_token: str = None,
878
cookies: Cookies = None,
879
headers: dict = None,
880
**kwargs
827
881
) -> AsyncIterator:
828
if cls._expires is not None and (cls._expires - 60*10) < time.time():
882
if cls._expires is not None and (cls._expires - 60 * 10) < time.time():
829
883
cls._headers = cls._api_key = None
830
884
if cls._headers is None or headers is not None:
831
885
cls._headers = {} if headers is None else headers
@@ -858,6 +912,7 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
858
912
async def nodriver_auth(cls, proxy: str = None):
859
913
async with get_nodriver_session(proxy=proxy) as browser:
860
914
page = await browser.get(cls.url)
915
861
916
def on_request(event: nodriver.cdp.network.RequestWillBeSent, page=None):
862
917
if event.request.url == start_url or event.request.url.startswith(conversation_url):
863
918
if cls.request_config.headers is None:
@@ -866,9 +921,10 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
866
921
cls.request_config.headers[key.lower()] = value
867
922
elif event.request.url in (backend_url, backend_anon_url):
868
923
if "OpenAI-Sentinel-Proof-Token" in event.request.headers:
869
cls.request_config.proof_token = json.loads(base64.b64decode(
870
event.request.headers["OpenAI-Sentinel-Proof-Token"].split("gAAAAAB", 1)[-1].split("~")[0].encode()
871
).decode())
924
cls.request_config.proof_token = json.loads(base64.b64decode(
925
event.request.headers["OpenAI-Sentinel-Proof-Token"].split("gAAAAAB", 1)[-1].split("~")[
926
0].encode()
927
).decode())
872
928
if "OpenAI-Sentinel-Turnstile-Token" in event.request.headers:
873
929
cls.request_config.turnstile_token = event.request.headers["OpenAI-Sentinel-Turnstile-Token"]
874
930
if "Authorization" in event.request.headers:
@@ -881,6 +937,7 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
881
937
arkBody=event.request.post_data,
882
938
userAgent=event.request.headers.get("User-Agent")
883
939
)
940
884
941
await page.send(nodriver.cdp.network.enable())
885
942
page.add_handler(nodriver.cdp.network.RequestWillBeSent, on_request)
886
943
await page.reload()
@@ -912,7 +969,7 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
912
969
if cls._api_key is not None or not cls.needs_auth:
913
970
break
914
971
await asyncio.sleep(1)
915
debug.log(f"OpenaiChat: Access token: {'False' if cls._api_key is None else cls._api_key[:12]+'...'}")
972
debug.log(f"OpenaiChat: Access token: {'False' if cls._api_key is None else cls._api_key[:12] + '...'}")
916
973
while True:
917
974
if cls.request_config.proof_token:
918
975
break
@@ -970,11 +1027,14 @@ class OpenaiChat(AsyncAuthedProvider, ProviderModelMixin):
970
1027
if cls._cookies:
971
1028
cls._headers["cookie"] = format_cookies(cls._cookies)
972
1029
1030
973
1031
class Conversation(JsonConversation):
974
1032
"""
975
1033
Class to encapsulate response fields.
976
1034
"""
977
def __init__(self, conversation_id: str = None, message_id: str = None, user_id: str = None, finish_reason: str = None, parent_message_id: str = None, is_thinking: bool = False):
1035
1036
def __init__(self, conversation_id: str = None, message_id: str = None, user_id: str = None,
1037
finish_reason: str = None, parent_message_id: str = None, is_thinking: bool = False):
978
1038
self.conversation_id = conversation_id
979
1039
self.message_id = message_id
980
1040
self.finish_reason = finish_reason
@@ -987,8 +1047,9 @@ class Conversation(JsonConversation):
987
1047
self.prompt = None
988
1048
self.generated_images: ImagePreview = None
989
1049
1050
990
1051
def get_cookies(
991
urls: Optional[Iterator[str]] = None
1052
urls: Optional[Iterator[str]] = None
992
1053
) -> Generator[Dict, Dict, Dict[str, str]]:
993
1054
params = {}
994
1055
if urls is not None:
@@ -1000,6 +1061,7 @@ def get_cookies(
1000
1061
json = yield cmd_dict
1001
1062
return {c["name"]: c["value"] for c in json['cookies']} if 'cookies' in json else {}
1002
1063
1064
1003
1065
class OpenAISources(ResponseType):
1004
1066
list: List[Dict[str, str]]
1005
1067
@@ -1025,7 +1087,7 @@ class OpenAISources(ResponseType):
1025
1087
if existing_source and idx is not None:
1026
1088
self.list[idx] = source
1027
1089
return
1028
1090
1029
1091
existing_source, idx = self.find_by_url(source["url"])
1030
1092
if existing_source and idx is not None:
1031
1093
self.list[idx] = source
@@ -1038,53 +1100,54 @@ class OpenAISources(ResponseType):
1038
1100
if not self.list:
1039
1101
return ""
1040
1102
return "\n\n\n\n" + ("\n>\n".join([
1041
f"> [{idx+1}] {format_link(link['url'], link.get('title', ''))}"
1103
f"> [{idx + 1}] {format_link(link['url'], link.get('title', ''))}"
1042
1104
for idx, link in enumerate(self.list)
1043
1105
]))
1044
1045
def get_ref_info(self, source: Dict[str, str]) -> dict[str, str|int] | None:
1106
1107
def get_ref_info(self, source: Dict[str, str]) -> dict[str, str | int] | None:
1046
1108
ref_index = source.get("ref_id", {}).get("ref_index", None)
1047
1109
ref_type = source.get("ref_id", {}).get("ref_type", None)
1048
1110
if isinstance(ref_index, int):
1049
1111
return {
1050
"ref_index": ref_index,
1112
"ref_index": ref_index,
1051
1113
"ref_type": ref_type,
1052
1114
}
1053
1115
1054
1116
for ref_info in source.get('refs') or []:
1055
1117
ref_index = ref_info.get("ref_index", None)
1056
1118
ref_type = ref_info.get("ref_type", None)
1057
1119
if isinstance(ref_index, int):
1058
1120
return {
1059
"ref_index": ref_index,
1121
"ref_index": ref_index,
1060
1122
"ref_type": ref_type,
1061
1123
}
1062
1124
1063
1125
return None
1064
1126
1065
def find_by_ref_info(self, ref_info: dict[str, str|int]):
1127
def find_by_ref_info(self, ref_info: dict[str, str | int]):
1066
1128
for idx, source in enumerate(self.list):
1067
1129
source_ref_info = self.get_ref_info(source)
1068
if (source_ref_info and
1069
source_ref_info["ref_index"] == ref_info["ref_index"] and
1070
source_ref_info["ref_type"] == ref_info["ref_type"]):
1071
return source, idx
1130
if (source_ref_info and
1131
source_ref_info["ref_index"] == ref_info["ref_index"] and
1132
source_ref_info["ref_type"] == ref_info["ref_type"]):
1133
return source, idx
1072
1134
1073
1135
return None, None
1074
1136
1075
1137
def find_by_url(self, url: str):
1076
1138
for idx, source in enumerate(self.list):
1077
1139
if source["url"] == url:
1078
1140
return source, idx
1079
return None, None
1141
return None, None
1080
1142
1081
def get_index(self, ref_info: dict[str, str|int]) -> int | None:
1143
def get_index(self, ref_info: dict[str, str | int]) -> int | None:
1082
1144
_, index = self.find_by_ref_info(ref_info)
1083
1145
if index is not None:
1084
return index
1146
return index
1085
1147
1086
1148
return None
1087
1149
1150
1088
1151
class ContentReferences:
1089
1152
def __init__(self) -> None:
1090
1153
self.list: List[Dict[str, Any]] = []
@@ -1098,16 +1161,16 @@ class ContentReferences:
1098
1161
1099
1162
self.list[idx] = {**self.list[idx], **reference_part}
1100
1163
1101
def update_reference(self, idx: int, operation: str, field: str, value: Any, ref_idx = None) -> None:
1164
def update_reference(self, idx: int, operation: str, field: str, value: Any, ref_idx=None) -> None:
1102
1165
while len(self.list) <= idx:
1103
1166
self.list.append({})
1104
1167
1105
1168
if operation == "append" or operation == "add":
1106
1169
if not isinstance(self.list[idx].get(field, None), list):
1107
1170
self.list[idx][field] = []
1108
1171
if isinstance(value, list):
1109
1172
self.list[idx][field].extend(value)
1110
else:
1173
else:
1111
1174
self.list[idx][field].append(value)
1112
1175
1113
1176
if operation == "replace" and ref_idx is not None:
@@ -1123,10 +1186,10 @@ class ContentReferences:
1123
1186
self.list[idx][field] = value
1124
1187
1125
1188
def get_ref_info(
1126
self,
1127
source: Dict[str, str],
1128
target_ref_info: Dict[str, Union[str, int]]
1129
) -> dict[str, str|int] | None:
1189
self,
1190
source: Dict[str, str],
1191
target_ref_info: Dict[str, Union[str, int]]
1192
) -> dict[str, str | int] | None:
1130
1193
for idx, ref_info in enumerate(source.get("refs", [])) or []:
1131
1194
if not isinstance(ref_info, dict):
1132
1195
continue
@@ -1134,11 +1197,11 @@ class ContentReferences:
1134
1197
ref_index = ref_info.get("ref_index", None)
1135
1198
ref_type = ref_info.get("ref_type", None)
1136
1199
if isinstance(ref_index, int) and isinstance(ref_type, str):
1137
if (not target_ref_info or
1138
(target_ref_info["ref_index"] == ref_index and
1139
target_ref_info["ref_type"] == ref_type)):
1200
if (not target_ref_info or
1201
(target_ref_info["ref_index"] == ref_index and
1202
target_ref_info["ref_type"] == ref_type)):
1140
1203
return {
1141
"ref_index": ref_index,
1204
"ref_index": ref_index,
1142
1205
"ref_type": ref_type,
1143
1206
"idx": idx
1144
1207
}
@@ -1149,9 +1212,9 @@ class ContentReferences:
1149
1212
for reference in self.list:
1150
1213
reference_ref_info = self.get_ref_info(reference, ref_info)
1151
1214
1152
if (not reference_ref_info or
1153
reference_ref_info["ref_index"] != ref_info["ref_index"] or
1154
reference_ref_info["ref_type"] != ref_info["ref_type"]):
1215
if (not reference_ref_info or
1216
reference_ref_info["ref_index"] != ref_info["ref_index"] or
1217
reference_ref_info["ref_type"] != ref_info["ref_type"]):
1155
1218
continue
1156
1219
1157
1220
if ref_info["ref_type"] != "image":
@@ -9,6 +9,8 @@ from pathlib import Path
9
9
from typing import Optional
10
10
from urllib.parse import urlparse
11
11
12
import requests
13
12
14
try:
13
15
from PIL import Image, ImageOps
14
16
has_requirements = True
@@ -383,6 +385,13 @@ def to_bytes(image: ImageType) -> bytes:
383
385
return Path(path).read_bytes()
384
386
else:
385
387
raise FileNotFoundError(f"File not found: {path}")
388
else:
389
resp = requests.get(image, headers={
390
"User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/137.0.0.0 Safari/537.36 Edg/137.0.0.0",
391
})
392
if resp.ok and is_accepted_format(resp.content):
393
return resp.content
394
raise ValueError("Invalid image url. Expected bytes, str, or PIL Image.")
386
395
else:
387
396
raise ValueError("Invalid image format. Expected bytes, str, or PIL Image.")
388
397
elif isinstance(image, Image.Image):
@@ -31,17 +31,21 @@ class StreamResponse(ClientResponse):
31
31
except json.JSONDecodeError:
32
32
continue
33
33
34
class StreamSession():
34
class StreamSession:
35
35
def __init__(
36
36
self,
37
headers: dict = {},
37
headers=None,
38
38
timeout: int = None,
39
39
connector: BaseConnector = None,
40
40
proxy: str = None,
41
proxies: dict = {},
41
proxies=None,
42
42
impersonate = None,
43
43
**kwargs
44
44
):
45
if proxies is None:
46
proxies = {}
47
if headers is None:
48
headers = {}
45
49
if impersonate:
46
50
headers = {
47
51
**DEFAULT_HEADERS,
@@ -49,7 +53,7 @@ class StreamSession():
49
53
}
50
54
connect = None
51
55
if isinstance(timeout, tuple):
52
connect, timeout = timeout;
56
connect, timeout = timeout
53
57
if timeout is not None:
54
58
timeout = ClientTimeout(timeout, connect)
55
59
if proxy is None: