site/backend/README.md

Workflow:

• should work with flask run locally and dockerfile build
• local dev
• local / venv pip install
• record versionless pips manually here
• pip freeze snapshots into project requirements
• docker build then copies frozen requirements

pip:

mariadb may take extra work: gcc, libmariadb-dev

pip install flask
pip install python-dotenv
pip install flask-wtf
pip install flask-sqlalchemy
pip install flask-migrate
pip install flask-login
pip install email-validator
pip install pydenticon
pip install flask-mail
pip install pyjwt
Prod only, require sys packages:
pip install mariadb
pip install uwsgi
...

Freeze/requirements.txt. Better to audit this inside python:3-bookworm-slim container.

pip freeze > requirements.txt

db cheat:

After db schema change:

flask db migrate -m "add users table"
flask db upgrade

Dump data if db in good state:

flask db downgrade base
flask db upgrade

Full reset or maria init:

sql:
drop table users;
drop table posts;
rm app.db
rm -r migrations
flask db init
flask db migrate
flask db upgrade

build notes:

Dockerfile needs dockerignore or preferably explicitly defined copies for:

• app
• config
• project dir
• requirements
• not dotflaskenv, vars set with dockerfile

docker setup:

• py logger handler StreamHandler is buggy to the point of being useless. Use logfile and link to stdout in Docker build

notes:

• compose has entry that overrides flask with uwsgi for prod
• miminal environment vars come through project env, pass through compose
• no dotenv here, dotflaskenv goes into image
• keep env untracked but templated, dotflaskenv is tracked and public