Name: trln-blacklight-quickstart
Owner: NCSU Libraries
Description: Quickstart Blacklight for TRLN investigation
Created: 2016-02-19 21:41:58.0
Updated: 2016-05-11 18:36:08.0
Pushed: 2016-08-02 14:38:56.0
Homepage: null
Size: 73
Language: Ruby
GitHub Committers
| User | Most Recent Commit | # Commits |
|---|
Other Committers
| User | Most Recent Commit | # Commits |
|---|
This repository contains a Rails application based on Blacklight 6, along with tools to populate a sample index. This is pre-alpha software!
This test was originally developed using Vagrant
and Puppet. The goal was to allow quickly spinning
up an instance of Blacklight for testing, with an eye to automating the
deployment of a potential Blacklight-fronted shared Solr index for TRLN. Most of what comes below assumes a Red Hat Enterprise Linux (RHEL) style environment, e.g. RHEL 6 or 7, or Centos 6/7. Vagrant box used was puppetlabs/centos-7.0.64-puppet, for reference.
This is currently an all-in-one quickstart, and doesn't cleanly separate out all of the components that would be involved in a production-ready deployment.
In particular, it leans on the chruby and ruby-install utilities to manage
ruby versions. We are investigating both JRuby and “standard” Ruby (aka MRI).
Currently, using MRI is recommended. We would like to explore JRuby as it
offers deployment and performance benefits (if it works).
Follow the instructions for installation for both projects, then run
by-install jruby
by-install ruby
will download and install the latest JRuby and MRI versions. Select an implementation with chruby [j]ruby (note you need to start a new shell session to pick up the tools.
System ruby may work, but has not been tested.
Java 8+
puppet installs openjdk 8, Oracle JDK will also work
Various RPMS
git
gcc
gcc-c++
libxml2, libxml2-devel
sqlite, sqlite-devel
Setup
This is a Rails application, so probably you should read up on that.
The init.sh (Bash) script will attempt to do as much of the initial setup as
it can, calling bundle install and performing some other tasks such as
creating the data directory for the default Solr core, downloading a suitable
version of Solr (into `$(pwd)/solr-install/solr-$SOLR_VERSION).
The script will also attempt to create the initial database. (bundle rake
db:migrate)
If any of this fails, well, it's early days yet. The script should tell most of the tale of what it's trying to do.
Blacklight uses a database to store search history and bookmarks for users. In this environment, we just have it use sqlite (a file-based database). This will probably buckle under any serious load testing =)
None yet.
We have had some success using traject to
do some very simple indexing of NCSU's MARC records. You will find a sample
traject configuration file in the config/traject subdirectory, which you can
copy and customize to your needs. Follow the comments in that file to create a copy of the traject configuration; I recommend keeping it in the same directory.
Next, edit config/trln.yml and make sure that development => traject =>
config_files has the right value, paying attention to the comments in that
file.
Next, find a middlin'-sized file full of MARC records (say, 100 or so) from your institution (traject will autodetect format and encoding), and execute our custom rake task:
will execute traject with your configuration file and attempt to insert the contents into the solr index. Traject's output will go to the console.
its authors, traject scales much better if you use JRuby to run it (`chruby
y`), but this probably won't matter for small datasets.
unning
ming you have all the above set up correctly, and have put the right
ion of ruby into your environment, start solr using either the solr_wrapper
`bundle exec solr_wrapper -c blacklight-core -d ./solr/conf`) or manually
ute `bundle rails server` (or `bundle rails s`) to start rails listening on
3000 (the Vagrantfile forwards this to 8088 on the host machine).
ut Why Not Use `solr_wrapper` intead of all this weird stuff in `lib/tasks/trln.rake` ?
klight tries to use `solr_wrapper`, a gem developed by experienced Blacklight / Ruby / Solr folks. But it turned out to be hard for me to figure out what it was doing and how it worked, so I decided to replicate a lot of what it does (no doubt, poorly) as part of an exercise in figuring out some of the Rails ecosystem.