Spaces:
Paused
Paused
Commit
·
840a4e4
1
Parent(s):
801882a
trying again
Browse files- main/main.py +33 -5
main/main.py
CHANGED
|
@@ -13,7 +13,16 @@ from huggingface_hub.hf_api import HfApi
|
|
| 13 |
logging.basicConfig(level=logging.INFO)
|
| 14 |
logger = logging.getLogger(__name__)
|
| 15 |
|
| 16 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 17 |
|
| 18 |
# Global variable to store the LLM instance
|
| 19 |
llm_instance = None
|
|
@@ -37,6 +46,20 @@ class GenerateRequest(BaseModel):
|
|
| 37 |
return_as_token_ids: bool = False
|
| 38 |
stream: bool = False
|
| 39 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 40 |
@app.post("/initialize")
|
| 41 |
async def initialize_model(request: InitializeRequest):
|
| 42 |
"""
|
|
@@ -167,12 +190,17 @@ async def health_check():
|
|
| 167 |
def main():
|
| 168 |
# Load environment variables or configuration here
|
| 169 |
host = os.getenv("LLM_ENGINE_HOST", "0.0.0.0")
|
| 170 |
-
port = int(os.getenv("LLM_ENGINE_PORT", "
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 171 |
|
| 172 |
# Start the server
|
| 173 |
-
hf_a = HfApi()
|
| 174 |
-
space_info = hf_a.space_info("TeamGenKI/LLM-Engine")
|
| 175 |
-
logger.warning(f"Exposing URL: {space_info}")
|
| 176 |
uvicorn.run(
|
| 177 |
app,
|
| 178 |
host=host,
|
|
|
|
| 13 |
logging.basicConfig(level=logging.INFO)
|
| 14 |
logger = logging.getLogger(__name__)
|
| 15 |
|
| 16 |
+
# Initialize FastAPI with root path for Spaces
|
| 17 |
+
app = FastAPI(
|
| 18 |
+
title="LLM Engine Service",
|
| 19 |
+
# This is crucial for Hugging Face Spaces
|
| 20 |
+
root_path="/",
|
| 21 |
+
# Add OpenAPI configs
|
| 22 |
+
openapi_url="/api/openapi.json",
|
| 23 |
+
docs_url="/api/docs",
|
| 24 |
+
redoc_url="/api/redoc"
|
| 25 |
+
)
|
| 26 |
|
| 27 |
# Global variable to store the LLM instance
|
| 28 |
llm_instance = None
|
|
|
|
| 46 |
return_as_token_ids: bool = False
|
| 47 |
stream: bool = False
|
| 48 |
|
| 49 |
+
@app.get("/")
|
| 50 |
+
async def root():
|
| 51 |
+
"""Root endpoint to verify service is running"""
|
| 52 |
+
space_url = "https://teamgenki-llm-engine.hf.space"
|
| 53 |
+
return {
|
| 54 |
+
"status": "running",
|
| 55 |
+
"service": "LLM Engine",
|
| 56 |
+
"endpoints": {
|
| 57 |
+
"initialize": f"{space_url}/initialize",
|
| 58 |
+
"generate": f"{space_url}/generate",
|
| 59 |
+
"health": f"{space_url}/health"
|
| 60 |
+
}
|
| 61 |
+
}
|
| 62 |
+
|
| 63 |
@app.post("/initialize")
|
| 64 |
async def initialize_model(request: InitializeRequest):
|
| 65 |
"""
|
|
|
|
| 190 |
def main():
|
| 191 |
# Load environment variables or configuration here
|
| 192 |
host = os.getenv("LLM_ENGINE_HOST", "0.0.0.0")
|
| 193 |
+
port = int(os.getenv("LLM_ENGINE_PORT", "7860")) # Changed to 7860 for Spaces
|
| 194 |
+
|
| 195 |
+
# Log the service URLs
|
| 196 |
+
space_url = "https://teamgenki-llm-engine.hf.space"
|
| 197 |
+
logger.info(f"Service will be available at: {space_url}")
|
| 198 |
+
logger.info(f"API endpoints:")
|
| 199 |
+
logger.info(f" Initialize: {space_url}/initialize")
|
| 200 |
+
logger.info(f" Generate: {space_url}/generate")
|
| 201 |
+
logger.info(f" Health: {space_url}/health")
|
| 202 |
|
| 203 |
# Start the server
|
|
|
|
|
|
|
|
|
|
| 204 |
uvicorn.run(
|
| 205 |
app,
|
| 206 |
host=host,
|