gpt4free/g4f/client/client.py
abc a107d3f2ff update default gpt-3.5-turbo models
by default, only OpenAI was enabled, more models enable for more flexibility.

new provider `Koala` added, to watch out as it could be unstable.
2024-04-12 17:06:54 +01:00

171 lines
6.3 KiB
Python

from __future__ import annotations
import time
import random
import string
from ..typing import Union, Iterator, Messages, ImageType
from ..providers.types import BaseProvider, ProviderType, FinishReason
from ..image import ImageResponse as ImageProviderResponse
from ..errors import NoImageResponseError
from .stubs import ChatCompletion, ChatCompletionChunk, Image, ImagesResponse
from .image_models import ImageModels
from .types import IterResponse, ImageProvider
from .types import Client as BaseClient
from .service import get_model_and_provider, get_last_provider
from .helper import find_stop, filter_json, filter_none
def iter_response(
response: iter[str],
stream: bool,
response_format: dict = None,
max_tokens: int = None,
stop: list = None
) -> IterResponse:
content = ""
finish_reason = None
completion_id = ''.join(random.choices(string.ascii_letters + string.digits, k=28))
for idx, chunk in enumerate(response):
if isinstance(chunk, FinishReason):
finish_reason = chunk.reason
break
content += str(chunk)
if max_tokens is not None and idx + 1 >= max_tokens:
finish_reason = "length"
first, content, chunk = find_stop(stop, content, chunk if stream else None)
if first != -1:
finish_reason = "stop"
if stream:
yield ChatCompletionChunk(chunk, None, completion_id, int(time.time()))
if finish_reason is not None:
break
finish_reason = "stop" if finish_reason is None else finish_reason
if stream:
yield ChatCompletionChunk(None, finish_reason, completion_id, int(time.time()))
else:
if response_format is not None and "type" in response_format:
if response_format["type"] == "json_object":
content = filter_json(content)
yield ChatCompletion(content, finish_reason, completion_id, int(time.time()))
def iter_append_model_and_provider(response: IterResponse) -> IterResponse:
last_provider = None
for chunk in response:
last_provider = get_last_provider(True) if last_provider is None else last_provider
chunk.model = last_provider.get("model")
chunk.provider = last_provider.get("name")
yield chunk
class Client(BaseClient):
def __init__(
self,
provider: ProviderType = None,
image_provider: ImageProvider = None,
**kwargs
) -> None:
super().__init__(**kwargs)
self.chat: Chat = Chat(self, provider)
self.images: Images = Images(self, image_provider)
class Completions():
def __init__(self, client: Client, provider: ProviderType = None):
self.client: Client = client
self.provider: ProviderType = provider
def create(
self,
messages: Messages,
model: str,
provider: ProviderType = None,
stream: bool = False,
proxy: str = None,
response_format: dict = None,
max_tokens: int = None,
stop: Union[list[str], str] = None,
api_key: str = None,
ignored : list[str] = None,
ignore_working: bool = False,
ignore_stream: bool = False,
**kwargs
) -> Union[ChatCompletion, Iterator[ChatCompletionChunk]]:
model, provider = get_model_and_provider(
model,
self.provider if provider is None else provider,
stream,
ignored,
ignore_working,
ignore_stream,
)
stop = [stop] if isinstance(stop, str) else stop
response = provider.create_completion(
model, messages,
stream=stream,
**filter_none(
proxy=self.client.get_proxy() if proxy is None else proxy,
max_tokens=max_tokens,
stop=stop,
api_key=self.client.api_key if api_key is None else api_key
),
**kwargs
)
response = iter_response(response, stream, response_format, max_tokens, stop)
response = iter_append_model_and_provider(response)
return response if stream else next(response)
class Chat():
completions: Completions
def __init__(self, client: Client, provider: ProviderType = None):
self.completions = Completions(client, provider)
def iter_image_response(response: Iterator) -> Union[ImagesResponse, None]:
for chunk in list(response):
if isinstance(chunk, ImageProviderResponse):
return ImagesResponse([Image(image) for image in chunk.get_list()])
def create_image(client: Client, provider: ProviderType, prompt: str, model: str = "", **kwargs) -> Iterator:
prompt = f"create a image with: {prompt}"
if provider.__name__ == "You":
kwargs["chat_mode"] = "create"
return provider.create_completion(
model,
[{"role": "user", "content": prompt}],
stream=True,
proxy=client.get_proxy(),
**kwargs
)
class Images():
def __init__(self, client: Client, provider: ImageProvider = None):
self.client: Client = client
self.provider: ImageProvider = provider
self.models: ImageModels = ImageModels(client)
def generate(self, prompt, model: str = None, **kwargs) -> ImagesResponse:
provider = self.models.get(model, self.provider)
if isinstance(provider, type) and issubclass(provider, BaseProvider):
response = create_image(self.client, provider, prompt, **kwargs)
else:
response = list(provider.create(prompt))
image = iter_image_response(response)
if image is None:
raise NoImageResponseError()
return image
def create_variation(self, image: ImageType, model: str = None, **kwargs):
provider = self.models.get(model, self.provider)
result = None
if isinstance(provider, type) and issubclass(provider, BaseProvider):
response = provider.create_completion(
"",
[{"role": "user", "content": "create a image like this"}],
True,
image=image,
proxy=self.client.get_proxy(),
**kwargs
)
result = iter_image_response(response)
if result is None:
raise NoImageResponseError()
return result