Service handling model placement
| k8s | ||
| model_service | ||
| .drone.yml | ||
| .gitignore | ||
| Dockerfile | ||
| README.md | ||
| requirements-dev.txt | ||
| requirements.txt | ||
Model Service
This service is a smart-storage for models used in various Ai algorithms (SVM and CNN) in the system. It also stores a "default pointer" to the Ai model intended to be used by the system.
The different Ai models can be accessed trough different endpoints.
This service requires a MinIO service and a relational database (possibly PostgreSQL). It does not initiate connection to other services.
Configure
- SENTRY_DSN: Sentry DSN. See https://docs.sentry.io/product/sentry-basics/dsn-explainer/
- RELEASE_ID: Release ID for Sentry.
- RELEASEMODE: Release mode for sentry.
- DATABASE_URI: URI connection string for the relational database. (e.g.:
postgresql://birbnetes:supersecret@model-postgres/model-service) - MINIO_ENDPOINT: MinIO connection configuration. See https://pypi.org/project/Flask-Minio/
- MINIO_ACCESS_KEY: MinIO connection configuration. See https://pypi.org/project/Flask-Minio/
- MINIO_SECRET_KEY: MinIO connection configuration. See https://pypi.org/project/Flask-Minio/
- MINIO_SVM_BUCKET_NAME: Name of the bucket to store SVM model files.
- MINIO_CNN_BUCKET_NAME: Name of the bucket to store CNN model files.
- MINIO_SECURE: MinIO connection configuration. See https://pypi.org/project/Flask-Minio/
- MINIO_REGION: MinIO connection configuration. See https://pypi.org/project/Flask-Minio/
- SECRET_KEY: Flask secret key. See https://flask.palletsprojects.com/en/2.0.x/config/#SECRET_KEY