Skip to content

Latest commit

 

History

History
242 lines (165 loc) · 7.88 KB

local.md

File metadata and controls

242 lines (165 loc) · 7.88 KB
Raw

Local

Dependencies required to run Orga Server

  • Python 3.7
  • Postgres
  • OpenSSL

For mac users

brew install postgresql

For debian-based linux users

sudo apt-get update
sudo apt-get install postgresql postgresql-contrib libssl-dev

Steps

Make sure you have the dependencies mentioned above installed before proceeding further.

  • Step 0 - Clone the Open Event Server repository (from the development branch) and cd into the directory.
git clone -b development https://github.com/fossasia/open-event-server.git
cd open-event-server

Note : If you want to contribute, first fork the original repository and clone the forked repository into your local machine followed by cd into the directory

git clone https://github.com/USERNAME/open-event-server.git
cd open-event-server
  • Step 1 - Install python3 requirements. You need to be present in the root directory of the project.

System Wide Installation

sudo -H pip3 install -r requirements.txt

hint: You may need to upgrade your pip version and install following packages if you encounter errors while installing the requirements.

Note: For Mac OS Sierra users, if you get an error that 'openssl/aes.h' could not be found when installing requirements.txt using pip, try the steps shown here - OSX openssl header error

Installation in Virtual Environment

You can use either pip or pipenv to install Open Event Server in a virtual environment.

Firstly, open a terminal and enter

# For linux users
sudo apt-get install python3-dev
sudo apt-get install libpq-dev
sudo apt-get install libffi6 libffi-dev

# For macOS users
brew install python@3
brew install libmagic

Using pip and virtualenv

Open a terminal and enter the following commands to setup a virtual environment

sudo apt-get install python3.7
virtualenv -p python3.7 venv
. venv/bin/activate

Now to install the dependencies using pip, type

pip3 install -r requirements.txt

Using pipenv

Using pipenv, you will not need to set up virtualenv. It will do it automatically for you

To setup a virtual environment and install the dependices, enter in a terminal

pipenv --python 3.7.3 install

Now to activate the virtual environemnt, type

pipenv shell
  • Step 2 - Create the database. For that we first open the psql shell. Go to the directory where your postgres file is stored.
# For linux users
sudo -u postgres psql

# For macOS users
psql -d postgres
  • When inside psql, create a user for open-event and then using the user create the database. Also, create a test database named opev_test for the test suites by dumping the oevent database into it. without this, the tests will not run locally.

For ease of development, you should create Postgres user with the same username as your OS account. If your OS login account is john, for example, you should create john user in Postgres. By this, you can skip entering password when using database.

CREATE USER open_event_user WITH PASSWORD 'opev_pass';
CREATE DATABASE oevent WITH OWNER open_event_user;
CREATE DATABASE opev_test WITH OWNER open_event_user;

  • Once the databases are created, exit the psql shell with \q followed by ENTER.

  • Step 3 - Create application environment variables.

cp .env.example .env

The URL is short, thanks to the resemble of Postgres user and OS user.

  • Step 4 - Start the postgres service.
sudo service postgresql restart

for mac users:

brew services restart postgresql
  • Step 5 - Create the tables. For that we will use create_db.py.
python3 create_db.py
# enter email and password
python3 manage.py db stamp head

Note 1: In case you made your own username and password in Step 2 are now getting FATAL: password authentication failed for user "john" , probable cause is non updation of .env file. To resolve it, open the .env file and update DATABASE_URL=postgresql://USERNAME:PASSWORD@127.0.0.1:5432/oevent and you are good to go.

Note 2: In case you are using Anaconda distribution for python, you may get an import error regarding celery.signals module. Please use the default python version while executing these steps in that case.

  • Step 6 - Start the application along with the needed services.
# Install and run redis
# For Ubuntu, Debian and alike
sudo apt-get install redis-server
# For Fedora, RedHat, CentOS
sudo dnf install redis

# For macOS
brew install redis

# Run Celery
# socketio has problems with celery "blocking" tasks
# also socketio is not used in a celery task so no problem to turn it off
INTEGRATE_SOCKETIO=false celery worker -A app.celery

# run app
python3 manage.py runserver
  • Step 7 - Rejoice. Go to localhost:5000 in your web browser to see the application live.

Flask-SocketIO development

Flask-SocketIO has been used in the project for displaying real-time notifications to the user. Although it's switched off by default. To integrate SocketIO you must set the INTEGRATE_SOCKETIO variable to true at bash.

export INTEGRATE_SOCKETIO="true"

The development server is the one that Flask ships with. It's based on Werkzeug and does not support WebSockets. If you try to run it, you'll get a RunTime error, something like: You need to use the eventlet server. . To test real-time notifications, you must use the Gunicorn web server with eventlet worker class.

If you've installed development requirements, you should have both gunicorn and eventlet installed. To run application on port 5000, execute the following instead of python3 manage.py runserver:

gunicorn app:app --worker-class eventlet -w 1 --bind 0.0.0.0:5000 --reload

-w specifies the number of worker classes to be used. --reload is used for development environments, so the server is restarted if any of the application python files change.

Now you should be able to access the website at localhost:5000.

Nginx

Gunicorn shouldn't be serving static files, it's supposed to run just the Flask application. You can use Nginx to serve static files and bypass other requests to the Gunicorn server, using it as a reverse proxy server. Proper configuration to enable proxying of WebSocket requests can be found in the Flask-SocketIO documentation: https://flask-socketio.readthedocs.io/en/latest/ (search for Nginx).

For Vagrant Machine

Doing the same for Vagrant machine requires some more configuration. If you're using the Vagrantfile provided in the repo, then you can check that the port forwarding is done as: 8001 -> 5000. So accessing the 8001 port in host machine will access the port 5000 in the guest (vagrant) machine. So in the guest machine, you need to run Nginx at port 5000 and gunicorn at some other port (let's assume port 5001).

map $http_upgrade $connection_upgrade {
    default upgrade;
    ''      close;
}

server {
    listen       5000;

    sendfile off;

    location /static {
        alias /vagrant/app/static;
    autoindex on;
    }

    location / {
    proxy_pass http://127.0.0.1:5001;

    proxy_redirect http://127.0.0.1:5001/ http://127.0.0.1:8001/;

    proxy_http_version 1.1;
    proxy_set_header Upgrade $http_upgrade;
    proxy_set_header Connection $connection_upgrade;

    break;
    }
}

You can directly use this configuration and put it inside sites-available (/etc/nginx/sites-available/nginx.conf) and create a symlink for it in sites-enabled (/etc/nginx/sites-enabled/nginx.conf).

Test the Nginx configuration and restart the Nginx server. Then run the Gunicorn server.

sudo service nginx testconfig # Should respond with "test is successful"
sudo service nginx restart
gunicorn app:app --worker-class eventlet -w 1 --bind 0.0.0.0:5001 --reload