summaryrefslogtreecommitdiffstats
path: root/g4f/client.py
diff options
context:
space:
mode:
Diffstat (limited to '')
-rw-r--r--g4f/client.py267
1 files changed, 267 insertions, 0 deletions
diff --git a/g4f/client.py b/g4f/client.py
new file mode 100644
index 00000000..117db375
--- /dev/null
+++ b/g4f/client.py
@@ -0,0 +1,267 @@
+from __future__ import annotations
+
+import re
+
+from .typing import Union, Generator, AsyncGenerator, Messages, ImageType
+from .base_provider import BaseProvider, ProviderType
+from .Provider.base_provider import AsyncGeneratorProvider
+from .image import ImageResponse as ImageProviderResponse
+from .Provider import BingCreateImages, Gemini, OpenaiChat
+from .errors import NoImageResponseError
+from . import get_model_and_provider
+
+ImageProvider = Union[BaseProvider, object]
+Proxies = Union[dict, str]
+
+def read_json(text: str) -> dict:
+ """
+ Parses JSON code block from a string.
+
+ Args:
+ text (str): A string containing a JSON code block.
+
+ Returns:
+ dict: A dictionary parsed from the JSON code block.
+ """
+ match = re.search(r"```(json|)\n(?P<code>[\S\s]+?)\n```", text)
+ if match:
+ return match.group("code")
+ return text
+
+def iter_response(
+ response: iter,
+ stream: bool,
+ response_format: dict = None,
+ max_tokens: int = None,
+ stop: list = None
+) -> Generator:
+ content = ""
+ idx = 1
+ chunk = None
+ finish_reason = "stop"
+ for idx, chunk in enumerate(response):
+ content += str(chunk)
+ if max_tokens is not None and idx > max_tokens:
+ finish_reason = "max_tokens"
+ break
+ first = -1
+ word = None
+ if stop is not None:
+ for word in list(stop):
+ first = content.find(word)
+ if first != -1:
+ content = content[:first]
+ break
+ if stream:
+ if first != -1:
+ first = chunk.find(word)
+ if first != -1:
+ chunk = chunk[:first]
+ else:
+ first = 0
+ yield ChatCompletionChunk([ChatCompletionDeltaChoice(ChatCompletionDelta(chunk))])
+ if first != -1:
+ break
+ if not stream:
+ if response_format is not None and "type" in response_format:
+ if response_format["type"] == "json_object":
+ response = read_json(response)
+ yield ChatCompletion([ChatCompletionChoice(ChatCompletionMessage(response, finish_reason))])
+
+async def aiter_response(
+ response: aiter,
+ stream: bool,
+ response_format: dict = None,
+ max_tokens: int = None,
+ stop: list = None
+) -> AsyncGenerator:
+ content = ""
+ try:
+ idx = 0
+ chunk = None
+ async for chunk in response:
+ content += str(chunk)
+ if max_tokens is not None and idx > max_tokens:
+ break
+ first = -1
+ word = None
+ if stop is not None:
+ for word in list(stop):
+ first = content.find(word)
+ if first != -1:
+ content = content[:first]
+ break
+ if stream:
+ if first != -1:
+ first = chunk.find(word)
+ if first != -1:
+ chunk = chunk[:first]
+ else:
+ first = 0
+ yield ChatCompletionChunk([ChatCompletionDeltaChoice(ChatCompletionDelta(chunk))])
+ if first != -1:
+ break
+ idx += 1
+ except:
+ ...
+ if not stream:
+ if response_format is not None and "type" in response_format:
+ if response_format["type"] == "json_object":
+ response = read_json(response)
+ yield ChatCompletion([ChatCompletionChoice(ChatCompletionMessage(response))])
+
+class Model():
+ def __getitem__(self, item):
+ return getattr(self, item)
+
+class ChatCompletion(Model):
+ def __init__(self, choices: list):
+ self.choices = choices
+
+class ChatCompletionChunk(Model):
+ def __init__(self, choices: list):
+ self.choices = choices
+
+class ChatCompletionChoice(Model):
+ def __init__(self, message: ChatCompletionMessage):
+ self.message = message
+
+class ChatCompletionMessage(Model):
+ def __init__(self, content: str, finish_reason: str):
+ self.content = content
+ self.finish_reason = finish_reason
+ self.index = 0
+ self.logprobs = None
+
+class ChatCompletionDelta(Model):
+ def __init__(self, content: str):
+ self.content = content
+
+class ChatCompletionDeltaChoice(Model):
+ def __init__(self, delta: ChatCompletionDelta):
+ self.delta = delta
+
+class Client():
+ proxies: Proxies = None
+ chat: Chat
+
+ def __init__(
+ self,
+ provider: ProviderType = None,
+ image_provider: ImageProvider = None,
+ proxies: Proxies = None,
+ **kwargs
+ ) -> None:
+ self.proxies: Proxies = proxies
+ self.images = Images(self, image_provider)
+ self.chat = Chat(self, provider)
+
+ def get_proxy(self) -> Union[str, None]:
+ if isinstance(self.proxies, str) or self.proxies is None:
+ return self.proxies
+ elif "all" in self.proxies:
+ return self.proxies["all"]
+ elif "https" in self.proxies:
+ return self.proxies["https"]
+ return None
+
+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,
+ response_format: dict = None,
+ max_tokens: int = None,
+ stop: list = None,
+ **kwargs
+ ) -> Union[dict, Generator]:
+ if max_tokens is not None:
+ kwargs["max_tokens"] = max_tokens
+ if stop:
+ kwargs["stop"] = list(stop)
+ model, provider = get_model_and_provider(
+ model,
+ self.provider if provider is None else provider,
+ stream,
+ **kwargs
+ )
+ response = provider.create_completion(model, messages, stream=stream, **kwargs)
+ if isinstance(provider, type) and issubclass(provider, AsyncGeneratorProvider):
+ response = iter_response(response, stream, response_format) # max_tokens, stop
+ else:
+ response = iter_response(response, stream, response_format, max_tokens, stop)
+ return response if stream else next(response)
+
+class Chat():
+ completions: Completions
+
+ def __init__(self, client: Client, provider: ProviderType = None):
+ self.completions = Completions(client, provider)
+
+class ImageModels():
+ gemini = Gemini
+ openai = OpenaiChat
+
+ def __init__(self, client: Client) -> None:
+ self.client = client
+ self.default = BingCreateImages(proxy=self.client.get_proxy())
+
+ def get(self, name: str) -> ImageProvider:
+ return getattr(self, name) if hasattr(self, name) else self.default
+
+class ImagesResponse(Model):
+ data: list[Image]
+
+ def __init__(self, data: list) -> None:
+ self.data = data
+
+class Image(Model):
+ url: str
+
+ def __init__(self, url: str) -> None:
+ self.url = url
+
+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):
+ provider = self.models.get(model) if model else self.provider or self.models.get(model)
+ if isinstance(provider, BaseProvider) or isinstance(provider, type) and issubclass(provider, BaseProvider):
+ prompt = f"create a image: {prompt}"
+ response = provider.create_completion(
+ "",
+ [{"role": "user", "content": prompt}],
+ True,
+ proxy=self.client.get_proxy()
+ )
+ else:
+ response = provider.create(prompt)
+
+ for chunk in response:
+ if isinstance(chunk, ImageProviderResponse):
+ return ImagesResponse([Image(image)for image in list(chunk.images)])
+ raise NoImageResponseError()
+
+ def create_variation(self, image: ImageType, model: str = None, **kwargs):
+ provider = self.models.get(model) if model else self.provider
+ if isinstance(provider, BaseProvider):
+ response = provider.create_completion(
+ "",
+ [{"role": "user", "content": "create a image like this"}],
+ True,
+ image=image,
+ proxy=self.client.get_proxy()
+ )
+ for chunk in response:
+ if isinstance(chunk, ImageProviderResponse):
+ return ImagesResponse([Image(image)for image in list(chunk.images)])
+ raise NoImageResponseError() \ No newline at end of file