/setup

for setting up new virtual envs

Credit to @ieshaw for this template. Some personal tweaks

Lightsail_Utils

Notes and utilities for spinning up AWS Lightsail instances. These are written for a fresh Lightsail linux instance.

SSH in to your instance

ssh -i KEY_LOCATION ec2-user@PUBLIC_IP

Git

Ensure git is installed

sudo yum install git -y

Then clone this repo

git clone https://github.com/rskuzma/setup.git

If you want to store your credentials (be careful, stroed as plain text) run this command

git config --global credential.helper store
git push

The push command will ask for your credentials, but now they will be stored.

To pull down a local copy of a remote non-master branch to work on

git checkout --track origin/branch_name

Python

To install Python 3.6 (needed for fastai)

sudo yum install -y python36u python36u-libs python36u-devel python36u-pipsudo pip install --upgrade pip

Requirements

Typical installs (note to self -- make bash script) Also, re-check this works for fastai

python36 -m venv env
source env/bin/activate
sudo yum install -y gcc
pip install git+https://github.com/fastai/fastai.git
pip install numpy sklearn pandas seaborn jupyter torch
pip install fastai

Jupyter Notebook

Praise goes to @rskuzma for figuring this out

  1. Ssh into lightsail instance using light sail privatekey

  2. $ jupyter notebook --no-browser --port=8888

  3. Open a new terminal

  4. ssh -i /PATH/TO/KEY/thisIsmyKey.pem -L 8000:localhost:8888 ec2-user@PUBLIC_IP_HERE

  5. If you get a “ WARNING: UNPROTECTED PRIVATE KEY FILE!” Then run this command chmod 400 path_to_private_key.pem

  6. Go to localhost:8000 in browser

  7. You’ll be at a jupyter notebook page requesting a token/login

  8. In your ssh window that’s now on your lights tail instance

  9. type $ jupyter notebook list

  10. Copy the part behind token=

  11. Paste that token into the jupyter notebook login page

  12. Make a password and save it to your password manager