Visualization

Service for visualisation of high dimensional for hydrosphere

Service for visualization of high dimensional for hydrosphere

DEPENDENCIES

DEBUG_ENV = bool(os.getenv("DEBUG_ENV", False))
APP_PORT = int(os.getenv("APP_PORT", 5000))
GRPC_PORT = os.getenv("GRPC_PORT", 5001)
GRPC_UI_ADDRESS = os.getenv("GRPC_UI_ADDRESS", "localhost:9090")
HS_CLUSTER_ADDRESS = os.getenv("HTTP_UI_ADDRESS", "http://localhost")
SECURE = os.getenv("SECURE", False)
MONGO_URL = os.getenv("MONGO_URL", "mongodb")
MONGO_PORT = int(os.getenv("MONGO_PORT", 27017))
MONGO_AUTH_DB = os.getenv("MONGO_AUTH_DB", "admin")
MONGO_USER = os.getenv("MONGO_USER")
MONGO_PASS = os.getenv("MONGO_PASS")
AWS_STORAGE_ENDPOINT = os.getenv('AWS_STORAGE_ENDPOINT', '')
AWS_REGION = os.getenv('AWS_REGION', '')
HYDRO_VIS_BUCKET_NAME = os.getenv('AWS_BUCKET', 'hydro-vis')

Assumptions:

  • Model must have in it’s contract ’embedding’ output
  • If model returns class prediction and confidence these fields should be named ‘class’ and ‘confidence’ respectively
  • Only data (embeddings) from requests will be visualized. Training data is used only for accurate transformation.

API

Task states:

hydro-visualization

Whole API description is available here

1.POST /visualization/plottable_embeddings/umap?model_version_id=2

transformer - manifold learning


 


 




 


To finish reading, please visit source site