A little base project to get up to speed with web development and deployment using Flask, HTMx and shell, with a simple model, view/blueprint, and template available.
You can:
- Clone this template repo:
git clone https://github.com/magnetrwn/flask-htmx-docker
. - Run the clone with:
docker-compose build && docker-compose up
. - Or just get the image with:
docker pull ghcr.io/magnetrwn/flask-htmx-docker:nightly
You can also run the clone locally by using the scripts in scripts/
. Use scripts/run-app-gunicorn.sh debug
to run a local Gunicorn instance, but make sure you have run scripts/setup-python-env.sh
at least once.
I wanted to expand a Flask/HTMx project I had found here, so I separated views, models and templates further, added db and log folders, and setup Docker for deployment.
Now, this structure can be repurposed for other simple Flask/HTMx web applications!
Currently, as you can see in requirements.txt
, the image installs the following using pip3 in a persistent virtual environment appenv
:
flask
flask-sqlalchemy
gunicorn
The db/
, log/
and appenv/
folders are persistent, and are represented as volumes /db
, /log
and /flask-htmx/appenv
in deployment.
Note: When using the Docker image, make sure to replace the default contents of the following directories:
# Replace these with your own:
/flask-htmx/app/models/
/flask-htmx/app/templates/
/flask-htmx/app/views/
# Include your new models and views here:
/flask-htmx/app/__init__.py
You can use the defaults to debug the image before starting Flask development.
To replace the contents of the image, you should use this Dockerfile syntax:
FROM ghcr.io/magnetrwn/flask-htmx-docker:nightly
# Replace the template app entirely
COPY my-app /flask-htmx/app
# Make sure to have variable 'app' available with Flask in 'app/__init__.py',
# as well as importing all models and views (check the file for clarity).
# Why? check note below.
Note: When starting Gunicorn, it searches for variable app
inside of module app
: you can find it in app/__init__.py
. This is why it also contains the blueprint registration logic, although this approach is only appropriate for compact web designs, so you should further separate __init__.py
moving on.
The default database used by this template is SQLite, and it is embedded for lightweight database usage, but you should integrate with larger database systems for heavier applications, since SQLite is not good for parallel performance: it will mutex one client at a time on the whole database file!
Licensed under MIT. Feel free to contribute if you want to!