Simple fullstack and monorepo PoC for data visualization, using Python/Flask, Node/Vue.js and Postgres, deployed as containers handling backend, frontend and database. Test it locally with docker-compose/k3s, running the individual container images or via scripts.

container images at hub.

Featuring:

• Backend:Python/Flask
  • Session-Cookie Authentication: flask_login
  • PyJWT
• Frontend: Node/Vue.js
  • Build tooling: Vite
  • HTTP requests: Axios
  • State Management: pinia
  • Validation: vee-validate, yup, “@vee-validate/yup”
  • Routing: vue-router
  • Visualization: vue-chartjs
• Database ORM: SQLAlchemy
  • postgres for dataframe storage
  • sqlite3 for session cache [?]

manually running the process

Scaffolding generated with [email protected]:

; npm create [email protected] frontend/ -- --template vue-ts
;
; npm install [email protected] [email protected] [email protected] [email protected] [email protected] [email protected]
;
; npm install -D '@types/node'
;
; npm run dev

run it as a standalone container so instances of postgres aren't polluted

Run the container process with the database

podman run -it -p=5432:5432 --name dbdrop -d \
    -e POSTGRES_USER=admin \
    -e POSTGRES_PASSWORD=Passw0rd \
    -v /mnt/ssd/dataStore/containers/database/pgad-pod/pgdata:/var/lib/postgresql/data:Z \
    docker.io/library/postgres:16.0

Connect psql to the process, create the schema and populate the database. The loop below reads all files under the database directory in alphanumeric order, so you can always tell that the reading respects precedence.

# Create schema with Raw SQL script then populate the database
#psql -U admin -d database_name -f=file.sql

for f in *.sql;
do
    psql -U admin -d dbdrop -f "$f"
done

manually running the process

Make sure to set the virtualenv:

; git clone [email protected]:isi23drop/vlusk-primer.git
; python3 -m venv venv

Prepare the environment:

; cd ./vlusk/backend/
; source ../venv/bin/activate
; pip3 install --upgrade pip
; pip3 install -r ./requirements.txt

Now that all dependencies are installed, just run each process from a different terminal:

1. database migration

flask --app app init-db

2. PS: now in flask 3.0 the FLASK_ENV is deprecated alongside the debug mode that goes back to the framework on the CLI call, hence the flag "--debug".

cd ./app/
flask --debug run

compose concentrates in orchestrating multiple containers in a single host. To do this with k8s, you would need kind, k3s (does not use virtualization) or similar. It was made to be compatible with other OCI runtimes, such as Podman, which was one of the first to enable rootless containers, and can be setup with compose using the Podman Service's systemd unit file for unix sockets.

The orchestration tool docker-compose supports Podman Service through the DOCKER_HOST environment variable. This makes it possible to run containers with podman but with the benefit of rootless.

Source the script and run it to run compose with podman:

; source ./scripts/podman-service-compose.sh

Kubernetes is a container orchestrator. To run this project on single host just like the Compose tool, you can use tools like kind or k3s (which don't use virtualization).