server setup

Provision a Server and Access via SSH

Installing Essentials

sudo apt update
sudo apt install software-properties-common
sudo add-apt-repository ppa:deadsnakes/ppa
sudo apt install build-essential zlib1g-dev libncurses5-dev libgdbm-dev libnss3-dev libssl-dev libreadline-dev libffi-dev libbz2-dev libsqlite3-dev wget git
git config --global user.name ""
git config --global user.email ""

Reference: https://linuxize.com/post/how-to-install-python-3-7-on-ubuntu-18-04/

Installing Pyenv

Instead of installing from source, we want to be able to use multiple versions of python. PyEnv is a great choice!

curl https://pyenv.run | bash
# Copy pyenv initialization code into ~/.bashrc
exec $SHELL

Reference https://github.com/pyenv/pyenv-installer

Installing Python

pyenv install 3.7.1
pyenv global 3.7.1

Managing environments

mkdir new-awesome-project
pyenv virtualenv 3.7.1 new-awesome-project
cd new-awesome-project
echo "new-awesome-project" > .python-version

Reference: https://medium.com/@martinlabs/my-python-development-workflow-with-pyenv-2bfbc03a15a1

Install Ta-lib

mkdir ta-lib
cd ta-lib
wget http://prdownloads.sourceforge.net/ta-lib/ta-lib-0.4.0-src.tar.gz
tar -xzf ta-lib-0.4.0-src.tar.gz
cd ta-lib
sudo ./configure
sudo make
sudo make install
pip install ta-lib

Deploying

On your server,

mkdir algoName
echo "virtualenv-name" > algoName/.python-version
mkdir algoName.git
touch algoName.git/hooks/post-receive
vim algoName.git/hooks/post-receive
chmod +x algoName.git/hooks/post-receive

Update post-receive file with the following:

#!/bin/sh
git --work-tree=/home/ubuntu/applications/algoName --git-dir=/home/ubuntu/applications/algoName.git checkout -f
cd /home/ubuntu/applications/algoName
pip install -r requirements.txt

On your computer

git init
echo ".env" > .gitignore
git add .
git commit -am "initial commit"
git remote add production ssh://ubuntu@ip/[path_to_application_dir].git
git push production master

Starting Your Algorithm

You can always start directly by just running python app.py, however the downside of this is if your algorithm crashes or your server reboots or you terminate your terminal session, your algorithm will not be running anymore. There are many ways to overcome this. What we will do is to use a production battled solution with a good balance for simplicity - PM2, a process manager that daemonizes the process of your running algorithm.

sudo apt install nodejs
sudo apt install npm
npm install pm2@latest -g
pm2 ecosystem

Edit your ecosystem.config.js

To start application

pm2 start ecosystem.config.js --env production
pm2 save

To view logs

pm2 logs name

Reference:

joshteng/server.md