mirror of
https://github.com/xtekky/gpt4free.git
synced 2024-12-26 12:41:56 +03:00
97 lines
3.5 KiB
Python
97 lines
3.5 KiB
Python
from __future__ import annotations
|
|
|
|
from aiohttp import ClientSession
|
|
from aiohttp.client_exceptions import ContentTypeError
|
|
|
|
from ...typing import AsyncResult, Messages
|
|
from ..base_provider import AsyncGeneratorProvider, ProviderModelMixin
|
|
from ..helper import format_prompt
|
|
import json
|
|
|
|
|
|
class NexraBing(AsyncGeneratorProvider, ProviderModelMixin):
|
|
label = "Nexra Bing"
|
|
url = "https://nexra.aryahcr.cc/documentation/bing/en"
|
|
api_endpoint = "https://nexra.aryahcr.cc/api/chat/complements"
|
|
working = True
|
|
supports_gpt_4 = False
|
|
supports_stream = False
|
|
|
|
default_model = 'Bing (Balanced)'
|
|
models = ['Bing (Balanced)', 'Bing (Creative)', 'Bing (Precise)']
|
|
|
|
model_aliases = {
|
|
"gpt-4": "Bing (Balanced)",
|
|
"gpt-4": "Bing (Creative)",
|
|
"gpt-4": "Bing (Precise)",
|
|
}
|
|
|
|
@classmethod
|
|
def get_model_and_style(cls, model: str) -> tuple[str, str]:
|
|
# Default to the default model if not found
|
|
model = cls.model_aliases.get(model, model)
|
|
if model not in cls.models:
|
|
model = cls.default_model
|
|
|
|
# Extract the base model and conversation style
|
|
base_model, conversation_style = model.split(' (')
|
|
conversation_style = conversation_style.rstrip(')')
|
|
return base_model, conversation_style
|
|
|
|
@classmethod
|
|
async def create_async_generator(
|
|
cls,
|
|
model: str,
|
|
messages: Messages,
|
|
proxy: str = None,
|
|
stream: bool = False,
|
|
markdown: bool = False,
|
|
**kwargs
|
|
) -> AsyncResult:
|
|
base_model, conversation_style = cls.get_model_and_style(model)
|
|
|
|
headers = {
|
|
"Content-Type": "application/json",
|
|
"origin": cls.url,
|
|
"referer": f"{cls.url}/chat",
|
|
}
|
|
async with ClientSession(headers=headers) as session:
|
|
prompt = format_prompt(messages)
|
|
data = {
|
|
"messages": [
|
|
{
|
|
"role": "user",
|
|
"content": prompt
|
|
}
|
|
],
|
|
"conversation_style": conversation_style,
|
|
"markdown": markdown,
|
|
"stream": stream,
|
|
"model": base_model
|
|
}
|
|
async with session.post(cls.api_endpoint, json=data, proxy=proxy) as response:
|
|
response.raise_for_status()
|
|
try:
|
|
# Read the entire response text
|
|
text_response = await response.text()
|
|
# Split the response on the separator character
|
|
segments = text_response.split('\x1e')
|
|
|
|
complete_message = ""
|
|
for segment in segments:
|
|
if not segment.strip():
|
|
continue
|
|
try:
|
|
response_data = json.loads(segment)
|
|
if response_data.get('message'):
|
|
complete_message = response_data['message']
|
|
if response_data.get('finish'):
|
|
break
|
|
except json.JSONDecodeError:
|
|
raise Exception(f"Failed to parse segment: {segment}")
|
|
|
|
# Yield the complete message
|
|
yield complete_message
|
|
except ContentTypeError:
|
|
raise Exception("Failed to parse response content type.")
|