-
Notifications
You must be signed in to change notification settings - Fork 4
Deployment Instructions
Step-by-step instructions for deploying any of the Simplified applications -- circulation manager, metadata wrangler, or library registry. These instructions are designed for use on an EC2 AMI, or on a personal computer (for doing development).
You can also choose to deploy a Docker image.
sudo yum install python3 git postgresql python-sqlalchemy python-pip nginx
sudo pip install virtualenv virtualenvwrapper
# These are necessary for installing the xmlsec module through pip.
sudo yum install libxml2-devel xmlsec1-devel xmlsec1-openssl-devel libtool-ltdl-devel
# These are necessary for installing python lxml through pip.
sudo yum install libxslt-devel gcc libxslt-python
# This is necessary for installing psycopg2 through pip
sudo yum install postgresql-devel
# Used by pillow to generate JPG thumbnails.
sudo yum install libjpeg libjpeg-devel
sudo apt-get install python3-dev git postgresql python-sqlalchemy python-pip nginx
sudo pip install virtualenv virtualenvwrapper
sudo apt-get install libxml2-dev libxslt-dev gcc python-libxml2 libxmlsec1-dev libxml2-dev
# This is necessary for installing psycopg2 through pip
sudo apt-get install postgresql-server-dev-12
# Used by pillow to generate JPG thumbnails.
sudo apt-get install libjpeg-dev
sudo apt-get install python-numpy
-
Install homebrew:
ruby -e "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/install)" -
Install the XCode command line tools when prompted.
-
Download and install Postgres.app from postgresapp.com.
-
Set up SSH keys using github's instructions: https://help.github.com/articles/generating-ssh-keys/
-
Install Python, pip, virtualenv, and Postgres executables
# install python 3, with pip brew install python3 # install Node and NPM brew install node brew install pkg-config libffi # install virtualenv pip install virtualenv brew install libxmlsec1 # Add postgres executables to PATH export PATH=$PATH:/Applications/Postgres.app/Contents/Versions/12/bin/
Make sure to add the export PATH line to your .bashrc file -- this will let you run psql from the command line in the future.
- Install additional dependencies
brew install libjpeg
The circulation manager requires Elasticsearch version 6. 6.8.10, for example, is known to work. Other components of the system don't need Elasticsearch at all.
Follow the instructions at Install Elasticsearch with RPM. You'll need to the official Elasticsearch repository and import its GPG key. Then you can run this command:
sudo yum install elasticsearch
elasticsearch-plugin install analysis-icu
Make sure Elasticsearch starts on bootup! Here's how to start it manually:
sudo systemctl start elasticsearch
sudo apt-get install elasticsearch
elasticsearch-plugin install analysis-icu
Make sure Elasticsearch starts on bootup! Here's how to start it manually:
sudo service elasticsearch start
brew cask install homebrew/cask-versions/adoptopenjdk8
brew install elasticsearch
elasticsearch-plugin install analysis-icu
Check http://localhost:9200 to make sure the Elasticsearch server is running.
You're probably checking out the circulation manager. Here's the command:
git clone git@github.com:NYPL-Simplified/circulation.git circulation
cd circulation
Make sure you have an SSH key associated with your Github account, or you won't be able to clone a repository over SSH.
If you're checking out the metadata wrangler, the command is the same but the repository is different.
git clone git@github.com:NYPL-Simplified/metadata_wrangler.git metadata
cd metadata
Both product repositories include a Git submodule for the Simplified-server-core project, which defines common things like the database schema. You'll need to initialize this module.
git submodule init
git submodule update
When you run git submodule update, the Simplified-server-core project will be cloned into the core directory.
The circulation manager includes a front-end administrative interface written in Node. To get this working, you'll need to link an additional project called circulation-web to the circulation manager.
The simplest way to install this project is to run npm install within the circulation/api/admin directory:
cd circulation/api/admin
npm install
If you plan to do development on the administrative interface, you'll need to clone its repository rather than installing the latest release. Here's how to do that:
git clone git@github.com:NYPL-Simplified/circulation-web.git circulation-web
cd circulation-web
npm link
cd ../circulation/api/admin
npm link simplified-circulation-web
On Linux, run this command to get into a Postgresql client:
$ sudo -u postgres psql
On Mac OS X, start the Postgresql server and a psql session by clicking on the elephant icon.
Within the psql session, run these commands:
CREATE DATABASE simplified_circulation_test;
CREATE DATABASE simplified_circulation_dev;
CREATE USER simplified with password '[password]';
grant all privileges on database simplified_circulation_dev to simplified;
CREATE USER simplified_test with password '[password]';
grant all privileges on database simplified_circulation_test to simplified_test;
--Add pgcrypto to any circulation manager databases.
\c simplified_circulation_dev
create extension pgcrypto;
\c simplified_circulation_test
create extension pgcrypto;The next step is to create a virtual environment. To create the virtual environment, use virtualenv:
$ cd circulation # Or metadata
python3 -m venv env
Then add the following lines to the bottom of env/bin/activate:
export SIMPLIFIED_PRODUCTION_DATABASE="postgres://simplified:[password]@localhost:5432/simplified_circulation_dev"
export SIMPLIFIED_TEST_DATABASE="postgres://simplified_test:[password]@localhost:5432/simplified_circulation_test"
These environment variables will point the application server to the correct database. All further configuration will be done through the administrative interface and stored in the database.
PYCHARM NOTE: If you're running the web application using PyCharm, editing the activate file may not do the job. You may need to define these two environment variables in your global Pycharm configuration.
Now, activate the virtual environment:
source env/bin/activate
pip install -r requirements.txt
On the metadata wrangler and the circulation manager, download the TextBlob corpora. This shouldn't be necessary on the circulation manager, but for the moment it is.
$ python -m textblob.download_corpora
There are lots of technologies you can use to get your server up and running. Before you do any of them, make sure you can run python app.py and visit http://localhost:6500/ to see the application server running.
Done? Great! Now you can use whatever tools your dev team likes best to get the app to the people. If you don't have a preference, you might like to use our Deployment Instructions for Nginx & uWSGI. If you run into any problems, please share them along with your solutions, so we can update the instructions.
If you use a different stack, we'd love for you to send us a setup walk-through with us so we can add it to the wiki.
Once you have the app server running, you can start configuring libraries and importing content, like any other use of the circulation manager "Configuring a Demonstration Circulation Manager" will walk you through this process.
In particular, you'll need to connect the circulation manager to your Elasticsearch server and set up an initial library.