zottelsheep / meds_cloud_7 Goto Github PK

• VM aufsetzen mit LibVirt
• VM aufsetzen mit LibVirt und Ansible
• VM aufsetzen mit CloudInit
  • Hostname sinnvoll setzen
  • SSH-Key automatisch setzen
• Alles mit Ansible umsetzen

• In Kubernetes einlesen
• Erstellen eines Dockercontainers mit Pythonskript (z.B. Hello World)
• In Python Flask ansehen und JSON-Schnittstellen verstehen

• Basic Design of Website
• Reactive Elements
• Banner and pretty and stuff and other stuff

• Basic hosting (Website renders on browser)
• Use proper WSGI-Server

The backend-endpoint should be configurable from within flask:

• Figure out how to use templating or another pattern to change javascript-variable from within flask
• Use an environment-variable to set endpoint
• Set Environment-Variable inside Dockerfile / from kubespecification

• Add-Picture button should be outside the picture-window
• Picture should span the picture-window completely (Corners should get cut off)
• Add resend request button

Frontend should notify the user if a classification was unsucessfull

Usage of Pytest to test:

• Flask Website (maybe?)
• Flask API ( /predict)
• Internal components (Unit-Tests)

• Download local k8s env:
  • https://kubernetes.io/docs/tasks/tools/
  • minikube
  • kind
• Install kubectl
• Build Backend-Container?
• Write deployment yaml
• Deploy Stuff

Deploy the application to a kubernetes cluster

• Define Deployment-Specification for frontend and backend
• Define replicas of frontend and backend deployment (3 should suffice for now)
• Define service for frontend and backend (only cluster-ip, do not publish for now) and connect to service via kubectl to test out connection (second link is an interactive tutorial)
  https://kubernetes.io/docs/concepts/services-networking/service/
  https://kubernetes.io/docs/tutorials/kubernetes-basics/expose/expose-intro/
• Define an ingress with routes for frontend and backend:
  • Frontend Route: '/'
  • Backend Route: '/api/