finSurvival

Credit Survival Analysis: Full stack, credit risk application using FastAPI for backend service, PostgreSQL as database, Docker & Kubernetes for orchestration, RabbitMQ for messaging, PySurvival, MLFlow for the training environment, Streamlit for Frontend (API consuming).

Interactive API documentation: Exposed Endpoints

Streamlit Frontend: Login

Streamlit Frontend: Registration/Sign Up

Streamlit Frontend: Details Submission

Streamlit Frontend: Details Submission Info

Streamlit Frontend: Report View & Menu

Streamlit Frontend: Logout

Features

• Full Docker integration (Docker based).
• Kubernetes Microservices.
• Python FastAPI backends.
• Streamlit Frontend
• RabbitMQ Messaging Queue
• PostgreSQL Database
• MLFlow
• pySurvival

How to Run the Project Locally (Minikube)

Prerequisites

• minikube installed

Steps

1. Adjust the auth k8s manifest assign appropriate first superuser email, corresponding password and postgres password. Apply the same to other manifests that require them.

2. Create a postgres database named finlytik. For local use I recommend a postgres docker instance using Postgres recipe with IP adjustments to your suit. Run with the default or your choosing of POSTGRES_PASSWORD, FIRST_SUPERUSER_PASSWORD

3. Start minikube by running the following command:

   minikube start
   

4. Enable ingress in minikube installation by running the following command:

   minikube addons enable ingress
   

5. Set your /etc/hosts file with the following: cluster-ip finlytik.com cluster-ip rabbitmq-manager.com

   Get the cluster-ip by running the following command:

   minikube ip
   

6. Run the rabbitmq k8s manifests by running the following command from the folder:

   kubectl apply -f .
   

7. Go to rabbitmq-manager.com from your browser and login using guest for both the username and password. Create queues named profile and detail on RabbitMQ fro the Queues tab.

8. Run the other services by running the following command from the respective manifest folders:

   kubectl apply -f .
   

9. Go to finlytik.com from your browser and explore the application.