2023-09-03 11:26:26 +03:00
|
|
|
from __future__ import annotations
|
|
|
|
|
2024-06-26 14:55:03 +03:00
|
|
|
import json, requests, re
|
2023-08-28 02:43:45 +03:00
|
|
|
|
2024-08-29 09:03:32 +03:00
|
|
|
from curl_cffi import requests as cf_reqs
|
|
|
|
from ..typing import CreateResult, Messages
|
2024-06-26 14:55:03 +03:00
|
|
|
from .base_provider import ProviderModelMixin, AbstractProvider
|
2024-08-29 09:03:32 +03:00
|
|
|
from .helper import format_prompt
|
2023-08-28 02:43:45 +03:00
|
|
|
|
2024-06-26 14:55:03 +03:00
|
|
|
class HuggingChat(AbstractProvider, ProviderModelMixin):
|
2024-08-29 09:03:32 +03:00
|
|
|
url = "https://huggingface.co/chat"
|
|
|
|
working = True
|
2024-06-26 14:55:03 +03:00
|
|
|
supports_stream = True
|
2024-08-29 09:03:32 +03:00
|
|
|
default_model = "meta-llama/Meta-Llama-3.1-70B-Instruct"
|
2024-09-12 21:35:31 +03:00
|
|
|
|
2024-01-23 07:02:14 +03:00
|
|
|
models = [
|
2024-09-12 21:43:28 +03:00
|
|
|
'meta-llama/Meta-Llama-3.1-70B-Instruct',
|
2024-09-12 17:55:39 +03:00
|
|
|
'CohereForAI/c4ai-command-r-plus-08-2024',
|
2024-09-25 08:01:39 +03:00
|
|
|
'Qwen/Qwen2.5-72B-Instruct',
|
2024-10-03 00:40:04 +03:00
|
|
|
'meta-llama/Llama-3.2-11B-Vision-Instruct',
|
2024-09-25 20:05:39 +03:00
|
|
|
'NousResearch/Hermes-3-Llama-3.1-8B',
|
|
|
|
'mistralai/Mistral-Nemo-Instruct-2407',
|
|
|
|
'microsoft/Phi-3.5-mini-instruct',
|
2024-01-23 07:02:14 +03:00
|
|
|
]
|
2024-08-31 09:47:39 +03:00
|
|
|
|
2024-04-23 18:34:42 +03:00
|
|
|
model_aliases = {
|
2024-08-29 09:03:32 +03:00
|
|
|
"llama-3.1-70b": "meta-llama/Meta-Llama-3.1-70B-Instruct",
|
2024-09-12 17:55:39 +03:00
|
|
|
"command-r-plus": "CohereForAI/c4ai-command-r-plus-08-2024",
|
2024-09-25 08:01:39 +03:00
|
|
|
"qwen-2-72b": "Qwen/Qwen2.5-72B-Instruct",
|
2024-10-03 00:40:04 +03:00
|
|
|
"llama-3.2-11b": "meta-llama/Llama-3.2-11B-Vision-Instruct",
|
2024-09-25 20:05:39 +03:00
|
|
|
"hermes-3": "NousResearch/Hermes-3-Llama-3.1-8B",
|
|
|
|
"mistral-nemo": "mistralai/Mistral-Nemo-Instruct-2407",
|
|
|
|
"phi-3.5-mini": "microsoft/Phi-3.5-mini-instruct",
|
2024-04-23 18:34:42 +03:00
|
|
|
}
|
2023-08-28 02:43:45 +03:00
|
|
|
|
2024-08-29 09:03:32 +03:00
|
|
|
@classmethod
|
|
|
|
def get_model(cls, model: str) -> str:
|
|
|
|
if model in cls.models:
|
|
|
|
return model
|
|
|
|
elif model in cls.model_aliases:
|
|
|
|
return cls.model_aliases[model]
|
|
|
|
else:
|
|
|
|
return cls.default_model
|
2024-08-31 09:47:39 +03:00
|
|
|
|
2024-03-25 23:06:51 +03:00
|
|
|
@classmethod
|
2024-06-26 14:55:03 +03:00
|
|
|
def create_completion(
|
2023-08-28 02:43:45 +03:00
|
|
|
cls,
|
|
|
|
model: str,
|
2023-10-09 14:33:20 +03:00
|
|
|
messages: Messages,
|
2024-06-26 14:55:03 +03:00
|
|
|
stream: bool,
|
2024-08-29 09:03:32 +03:00
|
|
|
**kwargs
|
|
|
|
) -> CreateResult:
|
|
|
|
model = cls.get_model(model)
|
2024-08-31 09:47:39 +03:00
|
|
|
|
2024-08-29 09:03:32 +03:00
|
|
|
if model in cls.models:
|
2024-08-16 23:25:38 +03:00
|
|
|
session = cf_reqs.Session()
|
|
|
|
session.headers = {
|
|
|
|
'accept': '*/*',
|
|
|
|
'accept-language': 'en',
|
|
|
|
'cache-control': 'no-cache',
|
|
|
|
'origin': 'https://huggingface.co',
|
|
|
|
'pragma': 'no-cache',
|
|
|
|
'priority': 'u=1, i',
|
|
|
|
'referer': 'https://huggingface.co/chat/',
|
|
|
|
'sec-ch-ua': '"Not)A;Brand";v="99", "Google Chrome";v="127", "Chromium";v="127"',
|
|
|
|
'sec-ch-ua-mobile': '?0',
|
2024-06-26 14:55:03 +03:00
|
|
|
'sec-ch-ua-platform': '"macOS"',
|
2024-08-16 23:25:38 +03:00
|
|
|
'sec-fetch-dest': 'empty',
|
|
|
|
'sec-fetch-mode': 'cors',
|
|
|
|
'sec-fetch-site': 'same-origin',
|
|
|
|
'user-agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/127.0.0.0 Safari/537.36',
|
2024-06-26 14:55:03 +03:00
|
|
|
}
|
2024-08-31 09:47:39 +03:00
|
|
|
|
|
|
|
print(model)
|
2024-06-26 14:55:03 +03:00
|
|
|
json_data = {
|
2024-08-16 23:25:38 +03:00
|
|
|
'model': model,
|
2024-06-26 14:55:03 +03:00
|
|
|
}
|
2024-08-31 09:47:39 +03:00
|
|
|
|
2024-08-16 23:25:38 +03:00
|
|
|
response = session.post('https://huggingface.co/chat/conversation', json=json_data)
|
2024-06-26 14:55:03 +03:00
|
|
|
conversationId = response.json()['conversationId']
|
2024-08-31 09:47:39 +03:00
|
|
|
|
2024-09-12 21:43:28 +03:00
|
|
|
response = session.get(f'https://huggingface.co/chat/conversation/{conversationId}/__data.json?x-sveltekit-invalidated=11',)
|
2024-08-31 09:47:39 +03:00
|
|
|
|
2024-08-16 23:25:38 +03:00
|
|
|
data: list = (response.json())["nodes"][1]["data"]
|
|
|
|
keys: list[int] = data[data[0]["messages"]]
|
|
|
|
message_keys: dict = data[keys[0]]
|
|
|
|
messageId: str = data[message_keys["id"]]
|
2024-06-26 14:55:03 +03:00
|
|
|
|
|
|
|
settings = {
|
2024-08-29 09:03:32 +03:00
|
|
|
"inputs": format_prompt(messages),
|
|
|
|
"id": messageId,
|
|
|
|
"is_retry": False,
|
|
|
|
"is_continue": False,
|
|
|
|
"web_search": False,
|
|
|
|
"tools": []
|
2024-06-26 14:55:03 +03:00
|
|
|
}
|
|
|
|
|
2024-08-16 23:25:38 +03:00
|
|
|
headers = {
|
|
|
|
'accept': '*/*',
|
|
|
|
'accept-language': 'en',
|
|
|
|
'cache-control': 'no-cache',
|
|
|
|
'origin': 'https://huggingface.co',
|
|
|
|
'pragma': 'no-cache',
|
|
|
|
'priority': 'u=1, i',
|
|
|
|
'referer': f'https://huggingface.co/chat/conversation/{conversationId}',
|
|
|
|
'sec-ch-ua': '"Not)A;Brand";v="99", "Google Chrome";v="127", "Chromium";v="127"',
|
|
|
|
'sec-ch-ua-mobile': '?0',
|
|
|
|
'sec-ch-ua-platform': '"macOS"',
|
|
|
|
'sec-fetch-dest': 'empty',
|
|
|
|
'sec-fetch-mode': 'cors',
|
|
|
|
'sec-fetch-site': 'same-origin',
|
|
|
|
'user-agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/127.0.0.0 Safari/537.36',
|
|
|
|
}
|
|
|
|
|
|
|
|
files = {
|
2024-08-29 09:03:32 +03:00
|
|
|
'data': (None, json.dumps(settings, separators=(',', ':'))),
|
2023-08-28 02:43:45 +03:00
|
|
|
}
|
2024-04-22 21:02:17 +03:00
|
|
|
|
2024-08-16 23:25:38 +03:00
|
|
|
response = requests.post(f'https://huggingface.co/chat/conversation/{conversationId}',
|
|
|
|
cookies=session.cookies,
|
|
|
|
headers=headers,
|
|
|
|
files=files,
|
2024-06-26 14:55:03 +03:00
|
|
|
)
|
2024-08-31 09:47:39 +03:00
|
|
|
|
2024-06-26 14:55:03 +03:00
|
|
|
first_token = True
|
|
|
|
for line in response.iter_lines():
|
|
|
|
line = json.loads(line)
|
|
|
|
|
|
|
|
if "type" not in line:
|
|
|
|
raise RuntimeError(f"Response: {line}")
|
|
|
|
|
|
|
|
elif line["type"] == "stream":
|
|
|
|
token = line["token"]
|
|
|
|
if first_token:
|
|
|
|
token = token.lstrip().replace('\u0000', '')
|
|
|
|
first_token = False
|
2024-06-29 17:00:50 +03:00
|
|
|
else:
|
|
|
|
token = token.replace('\u0000', '')
|
|
|
|
|
2024-08-29 09:03:32 +03:00
|
|
|
yield token
|
2024-08-31 09:47:39 +03:00
|
|
|
|
2024-06-26 14:55:03 +03:00
|
|
|
elif line["type"] == "finalAnswer":
|
2024-08-31 09:47:39 +03:00
|
|
|
break
|