/flask-docker

Flask + Docker + Nginx + Gunicorn + MySQL + Factory Method Pattern

Primary LanguagePython

Author: Facundo Padilla

Social networks:

Version:

  • 1.0:
    • 01/02/2021 (DD-MM-YYYY)
    • 02/01/2021 (MM-DD-YYYY)
    • 2021/02/01 (YYYY-MM-DDDD)

What is this?:

  • This Flask project is reusable and also an example of how to merge Flask, Docker, Nginx, Gunicorn, MySQL, new: Flask-RESTX, Factory Method design pattern, and other optional dependencies such as Dynaconf, Marshmallow, SQLAlchemy, Faker, PyMySQL, Pytest, etc... which are installed inside the virtual environment "env_flask".

How to use it?

  • Easy, if you have Docker Compose installed, just run the following command inside the project directory: docker-compose up

  • You can also use the commands created inside the "Makefile" file, simply by executing the "make" command; make [OPTION] - Example: make full-start

  • Once you execute any of these commands, it starts to create the containers to work, they are already linked so you should have no problems to get it to work

  • Once you have finished running and creating the containers, simply go to http://localhost:80/

Initialize application:

  • After you have executed the commands like docker-compose up, make full-start or whatever you have created in the Makefile or used, it automatically creates the tables and works without problems, the tables are created in the predefined database with the name "flask_api", if you want to change the name of this database just go to the file "docker-compose.yaml" and change the environment variable "MYSQL_DATABASE".

  • And that's it, once the tables are created, you only have to go to http://localhost:80 (Nginx)

"Hot-Reloading":

  • You could use Watchman to do hot-reaload, but it gives some problems/bugs, so you can directly work with the virtual environment "env_flask" and start working there to simulate a hot-reload, also, it is not recommended that a container contains integrated hot-realoading if it is going to be used for hot-reloading.

  • How to activate env_flask:

    • Linux / PowerShell:

      1. Run: pip3 install virtualenv && python3 -m virtualenv env_flask
      2. Enter directory ../env_flask/bin/activate and execute: source activate (in the terminal)
      3. Return to the path where run_debug.py is found, and run: pip install -r requirements.txt
      4. Run the following commands:
        • export FLASK_APP=run_debug.py
        • flask run -h 0.0.0.0 -p 5001(the host and port can be changed to the one you want) or
        • python run_debug.py

    • Windows:

      1. Run: pip3 install virtualenv && python3 -m virtualenv env_flask
      2. Enter directory ../env_flask/bin/activate and execute: activate in CMD
      3. Return to the path where run_debug.py is found, and run: pip install -r requirements.txt
      4. Run the following commands:
        • set FLASK_APP=run_debug.py
        • flask run -h 0.0.0.0 -p 5001 (the host and port can be changed to the one you want) or
        • python run_debug.py

ATTENTION: in the Dynaconf configuration file (settings.toml), when running the run_debug.py , the MySQL connection points to localhost and port 3307, if it is going to be uploaded to production, remove the "expose" option from the docker-file.yaml file.

  • Deactivate the virtual environment:

    • Execute deactivate in the terminal or CMD

Settings:

  • docker-compose.yaml:

    • MySQL (db):

      • MYSQL_USER: the user name you want to customize (the root user is default and cannot be deleted)

      • MYSQL_PASSWORD= the password of the user (not the root)

      • MYSQL_DATABASE= name of the database you want

      • MYSQL_ROOT_PASSWORD= root user password

      • More documentation: https://hub.docker.com/_/mysql

    • Flask (flask_app):

      • PYTHONBUFFERED= by default leave it set to 1, it is used to display Python logs.

      • ENVVAR_PREFIX_FOR_DYNACONF= the name of our module

      • ENVVAR_FOR_DYNACONF= file name with extension ".toml", Dynaconf supports several others: https://dynaconf.readthedocs.io/en/docs_223/guides/examples.html

      • FLASK_APP= the name of the Python file to be executed by the server

      • FLASK_RUN_HOST= the host address of the application, defaulting to 0.0.0.0

      • FLASK_DEBUG= WARNING!, 1 enables debug, 0 for production.

      • command: Gunicorn command to execute

      • More documentation: https://hub.docker.com/_/python

    • Nginx (nginx lol):

  • settings.toml (Dynaconf file):

    • [default]

      • SQLALCHEMY_TRACK_MODIFICATIONS = "False" (is a setting that is no longer used, leave it at false)

    • [development]:

      When you switch to debug mode, the db name becomes localhost and port 3307.

      • DEBUG = "True"
      • SQLALCHEMY_DATABASE_URI = "mysql+pymysql://root:password@localhost:port/db_name"

    • [production]:

      When in production, mostly in the docker, it keeps running this configuration

      • DEBUG = "False"
      • SQLALCHEMY_DATABASE_URI = "mysql+pymysql://root:password@db_container_name/db_name"

python run.py vs flask run:

  • The difference between using Python to execute "run.py" and using "flask run", is that when you have the file "__module__.py" in the project, Python automatically adds it to the "syspath", in the case of "flask run", this does not happen, even if there is the "__init__.py" and the "__module__.py" , so to avoid problems when working with flask run, in each __init__.py file the following code fragment is added:

      import sys
  sys.path.append(".")