summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
Diffstat (limited to '')
-rw-r--r--g4f/Provider/GizAI.py151
1 files changed, 151 insertions, 0 deletions
diff --git a/g4f/Provider/GizAI.py b/g4f/Provider/GizAI.py
new file mode 100644
index 00000000..127edc9e
--- /dev/null
+++ b/g4f/Provider/GizAI.py
@@ -0,0 +1,151 @@
+from __future__ import annotations
+
+import json
+from aiohttp import ClientSession
+
+from ..typing import AsyncResult, Messages
+from ..image import ImageResponse
+from .base_provider import AsyncGeneratorProvider, ProviderModelMixin
+from .helper import format_prompt
+
+class GizAI(AsyncGeneratorProvider, ProviderModelMixin):
+ url = "https://app.giz.ai/assistant/"
+ api_endpoint = "https://app.giz.ai/api/data/users/inferenceServer.infer"
+ working = True
+
+ supports_system_message = True
+ supports_message_history = True
+
+ # Chat models
+ default_model = 'chat-gemini-flash'
+ chat_models = [
+ default_model,
+ 'chat-gemini-pro',
+ 'chat-gpt4m',
+ 'chat-gpt4',
+ 'claude-sonnet',
+ 'claude-haiku',
+ 'llama-3-70b',
+ 'llama-3-8b',
+ 'mistral-large',
+ 'chat-o1-mini'
+ ]
+
+ # Image models
+ image_models = [
+ 'flux1',
+ 'sdxl',
+ 'sd',
+ 'sd35',
+ ]
+
+ models = [*chat_models, *image_models]
+
+ model_aliases = {
+ # Chat model aliases
+ "gemini-flash": "chat-gemini-flash",
+ "gemini-pro": "chat-gemini-pro",
+ "gpt-4o-mini": "chat-gpt4m",
+ "gpt-4o": "chat-gpt4",
+ "claude-3.5-sonnet": "claude-sonnet",
+ "claude-3-haiku": "claude-haiku",
+ "llama-3.1-70b": "llama-3-70b",
+ "llama-3.1-8b": "llama-3-8b",
+ "o1-mini": "chat-o1-mini",
+ # Image model aliases
+ "sd-1.5": "sd",
+ "sd-3.5": "sd35",
+ "flux-schnell": "flux1",
+ }
+
+ @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
+
+ @classmethod
+ def is_image_model(cls, model: str) -> bool:
+ return model in cls.image_models
+
+ @classmethod
+ async def create_async_generator(
+ cls,
+ model: str,
+ messages: Messages,
+ proxy: str = None,
+ **kwargs
+ ) -> AsyncResult:
+ model = cls.get_model(model)
+
+ headers = {
+ 'Accept': 'application/json, text/plain, */*',
+ 'Accept-Language': 'en-US,en;q=0.9',
+ 'Cache-Control': 'no-cache',
+ 'Connection': 'keep-alive',
+ 'Content-Type': 'application/json',
+ 'Origin': 'https://app.giz.ai',
+ 'Pragma': 'no-cache',
+ 'Sec-Fetch-Dest': 'empty',
+ 'Sec-Fetch-Mode': 'cors',
+ 'Sec-Fetch-Site': 'same-origin',
+ 'User-Agent': 'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/130.0.0.0 Safari/537.36',
+ 'sec-ch-ua': '"Not?A_Brand";v="99", "Chromium";v="130"',
+ 'sec-ch-ua-mobile': '?0',
+ 'sec-ch-ua-platform': '"Linux"'
+ }
+
+ async with ClientSession() as session:
+ if cls.is_image_model(model):
+ # Image generation
+ prompt = messages[-1]["content"]
+ data = {
+ "model": model,
+ "input": {
+ "width": "1024",
+ "height": "1024",
+ "steps": 4,
+ "output_format": "webp",
+ "batch_size": 1,
+ "mode": "plan",
+ "prompt": prompt
+ }
+ }
+ async with session.post(
+ cls.api_endpoint,
+ headers=headers,
+ data=json.dumps(data),
+ proxy=proxy
+ ) as response:
+ response.raise_for_status()
+ response_data = await response.json()
+ if response_data.get('status') == 'completed' and response_data.get('output'):
+ for url in response_data['output']:
+ yield ImageResponse(images=url, alt="Generated Image")
+ else:
+ # Chat completion
+ data = {
+ "model": model,
+ "input": {
+ "messages": [
+ {
+ "type": "human",
+ "content": format_prompt(messages)
+ }
+ ],
+ "mode": "plan"
+ },
+ "noStream": True
+ }
+ async with session.post(
+ cls.api_endpoint,
+ headers=headers,
+ data=json.dumps(data),
+ proxy=proxy
+ ) as response:
+ response.raise_for_status()
+ result = await response.json()
+ yield result.get('output', '')