Spaces:
Running
Running
:gem: [Feature] New HuggingchatStreamer: get_conversation_id
Browse files- constants/networks.py +3 -0
- networks/huggingchat_streamer.py +80 -0
constants/networks.py
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
REQUESTS_HEADERS = {
|
| 2 |
+
"User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/123.0.0.0 Safari/537.36"
|
| 3 |
+
}
|
networks/huggingchat_streamer.py
ADDED
|
@@ -0,0 +1,80 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import json
|
| 2 |
+
import re
|
| 3 |
+
import requests
|
| 4 |
+
|
| 5 |
+
|
| 6 |
+
from tclogger import logger
|
| 7 |
+
from transformers import AutoTokenizer
|
| 8 |
+
|
| 9 |
+
from constants.models import (
|
| 10 |
+
MODEL_MAP,
|
| 11 |
+
STOP_SEQUENCES_MAP,
|
| 12 |
+
TOKEN_LIMIT_MAP,
|
| 13 |
+
TOKEN_RESERVED,
|
| 14 |
+
)
|
| 15 |
+
from constants.envs import PROXIES
|
| 16 |
+
from constants.networks import REQUESTS_HEADERS
|
| 17 |
+
from messagers.message_outputer import OpenaiStreamOutputer
|
| 18 |
+
|
| 19 |
+
|
| 20 |
+
class HuggingchatStreamer:
|
| 21 |
+
def __init__(self, model: str):
|
| 22 |
+
if model in MODEL_MAP.keys():
|
| 23 |
+
self.model = model
|
| 24 |
+
else:
|
| 25 |
+
self.model = "mixtral-8x7b"
|
| 26 |
+
self.model_fullname = MODEL_MAP[self.model]
|
| 27 |
+
self.message_outputer = OpenaiStreamOutputer(model=self.model)
|
| 28 |
+
# export HF_ENDPOINT=https://hf-mirror.com
|
| 29 |
+
self.tokenizer = AutoTokenizer.from_pretrained(self.model_fullname)
|
| 30 |
+
|
| 31 |
+
def count_tokens(self, text):
|
| 32 |
+
tokens = self.tokenizer.encode(text)
|
| 33 |
+
token_count = len(tokens)
|
| 34 |
+
logger.note(f"Prompt Token Count: {token_count}")
|
| 35 |
+
return token_count
|
| 36 |
+
|
| 37 |
+
def get_conversation_id(self, preprompt: str = ""):
|
| 38 |
+
request_url = "https://huggingface.co/chat/conversation"
|
| 39 |
+
request_body = {
|
| 40 |
+
"model": self.model_fullname,
|
| 41 |
+
"preprompt": preprompt,
|
| 42 |
+
}
|
| 43 |
+
logger.note(f"> Conversation ID:", end=" ")
|
| 44 |
+
res = requests.post(
|
| 45 |
+
request_url,
|
| 46 |
+
headers=REQUESTS_HEADERS,
|
| 47 |
+
json=request_body,
|
| 48 |
+
proxies=PROXIES,
|
| 49 |
+
timeout=10,
|
| 50 |
+
)
|
| 51 |
+
if res.status_code == 200:
|
| 52 |
+
conversation_id = res.json()["conversationId"]
|
| 53 |
+
logger.success(f"[{conversation_id}]")
|
| 54 |
+
else:
|
| 55 |
+
logger.warn(f"[{res.status_code}]")
|
| 56 |
+
raise ValueError("Failed to get conversation ID!")
|
| 57 |
+
self.conversation_id = conversation_id
|
| 58 |
+
|
| 59 |
+
def chat_response(
|
| 60 |
+
self,
|
| 61 |
+
prompt: str = None,
|
| 62 |
+
temperature: float = 0.5,
|
| 63 |
+
top_p: float = 0.95,
|
| 64 |
+
max_new_tokens: int = None,
|
| 65 |
+
api_key: str = None,
|
| 66 |
+
use_cache: bool = False,
|
| 67 |
+
):
|
| 68 |
+
pass
|
| 69 |
+
|
| 70 |
+
def chat_return_dict(self, stream_response):
|
| 71 |
+
pass
|
| 72 |
+
|
| 73 |
+
def chat_return_generator(self, stream_response):
|
| 74 |
+
pass
|
| 75 |
+
|
| 76 |
+
|
| 77 |
+
if __name__ == "__main__":
|
| 78 |
+
streamer = HuggingchatStreamer(model="mixtral-8x7b")
|
| 79 |
+
conversation_id = streamer.get_conversation_id()
|
| 80 |
+
# python -m networks.huggingchat_streamer
|