Dryad installation (v0.0.2)
The Dryad application is made of a number of parts intended to keep it more flexible and to separate concerns so that parts can be replaced with new metadata and other engines to customize it. Some basic information about the project and architechiture is available at the Dash Website, but this document focuses on getting Dash up and running for development.
You'll need the following parts installed and configured on a (local) UI development server to do development on the full UI application. Don't worry, there are more detailed installation instructions in other sections below and this is meant to give an overview of the larger dependencies to configure.
- (Recommended) A ruby version manager such as rbenv or rvm
- The bare Dryad application cloned from github
- The stash repository cloned from github
You'll also need the following components installed either on the same server or on separate servers for all the application features to work:
- MySQL (with the database specified in the database.yml created and using utf8mb4 character set by default)
- SOLR (with a geoblacklight schema and core installed)
- A storage repository that supports SWORD will be needed to submit documents to the repository and even with SWORD support, the code may need some customization for others besides the Merritt repository.
- A DOI minting service such as EZID to mint DOIs.
The application also requires some means to log in outside of a development environment. You'd want to configure a log in method for each application tenant from these:
- Google login
- Shibboleth login
- ORCID login (coming soon and required for ORCID lookup in the metadata page)
Installing the code and a base config
Open a (bash) shell and type these commands inside a directory where you want to work with this code. These will clone the development code and an example config.
git clone https://github.com/CDL-Dryad/dryad git clone https://github.com/CDL-Dryad/stash
Your config files will be stored in a separate directory from your application. It can be handy to keep them apart from the application so that you can back them up or commit them to a private repository for configuration separate from the application. The application will need to have these configuration files symlinked into the application. To copy the example config to an external directory and symlink the files in using a bash shell, type these commands:
cd dryad ./symlink_config.sh
You should end up with a directory structure that looks like this one.
├── dryad-config (contains actual config files) ├── dryad | ├── config (with symlinks to dryad-config above) | └── dryad-config-example └── stash ├── stash-harvester ├── stash-merritt ├── stash-sword ├── stash-wrapper ├── stash_datacite ├── stash_discovery └── stash_engine
Most of the configuration can be left as default. Items to check before first launch:
- dryad-config/app_config.yml, particularly the ORCID key and secret
Installing MySQL and Solr
The procedure to install MySQL and Solr vary from one operating system to another, but this guide shows a way to configure it in Ubuntu linux:
# installing MySQL in an Ubuntu Linux distro, make note of the root password you set while installing sudo apt-get install mysql-server mysql-client libmysqlclient-dev # make sure MySQL is started sudo service mysql start # connect to mysql, note the <username> is probably root in a new installation, and the password is probably blank mysql -u <username> -p # if the above doesn't work, try sudo mysql -u root # create the dash database CREATE DATABASE dryad CHARACTER SET utf8mb4 COLLATE utf8mb4_unicode_ci; # add a user to the database CREATE USER 'travis'@'%'; # grant the user privileges GRANT ALL PRIVILEGES ON dryad . * TO 'travis'@'%'; FLUSH PRIVILEGES; # To exit the MySQL client, type *exit* or press ctrl-d
Now edit the dryad-config/config/database.yml file to fill in the dashuser and password you set above in the development environment for that configuration file.
Solr requires a Java runtime. Try java -version and if it says that "java can be found in the following packages" rather than giving you a version you probably need to install java with a command like sudo apt-get install default-jre .
Make sure Solr is working by going to http://localhost:8983. You should see a Solr admin page.
Verify Solr is set up correctly from the Admin UI:
You can then click the query sidebar tab and scroll down to the bottom of the form to submit a blank query. While the document will not return any results yet because there are no documents in SOLR, you should see it execute and you can verify that Solr queries are running.
(Optional, but recommended) Add a sample record to match the sample database record (see below).
Find the file dryad-config/sample_data/sample_record.json in the dryad-config repo. Open the file in a text editor, select all the text and copy it.
Getting the Rails application running
I'd strongly recommend installing rbenv for a local development asenvironment as a way to manage Ruby versions. Follow the installation instructions given on the rbenv site to install it, but make sure the
rbenv init command is run in every shell (e.g., add it to .bashrc). Install the Ruby build plugin to make it easy to install different Ruby versions as needed.
Note for OSX: Ensure the command-line XCode utilities are installed.
# make sure some basic libraries are installed that are probably required later (Ubuntu example) sudo apt-get install libxml2 libxml2-dev patch curl cd dryad rbenv install $(cat .ruby-version) # installs the ruby-version set in the .ruby-version file # update your rubygems version gem update --system # install bundler to handle gem dependencies gem install bundler # now install the gem libraries needed for the application bundle install # run the migrations to set up the database tables bundle exec rake db:migrate # start your rails server for local development rails s
If you want to view sample data, then insert a sample record into the database (recommended).
# connect to mysql, note the <username> is probably root in a new installation mysql -u <username> -p # Use the following two lines. USE dash; source ../dryad-config/sample_data/sample_record.sql; # To exit the MySQL client, type *exit* or press ctrl-d
To configure where the search enterface draws its data from, modify the dryad app config/blacklight.yml to change the endpoint for the development server. When running locally, the default server is development.
Testing basic functionality
Explore the datasets
Open your web browser and go to http://localhost:3000 to see the homepage.
The Explore Data link will allow you to search and view your dataset, if you imported a sample record.
Enter dataset metadata, upload files and preview the landing page
After you log in, you will be able to start entering metadata and uploading files for a dataset by clicking the My Datasets menu link.
Metadata entry, file uploading and landing page preview should be functional.
We have enabled submission to a SWORD-enabled Merritt repository, but have only implemented relevant parts of the SWORD specification and not every functionality in the specification has been implemented.
Next steps in configuration
Repository and identifier service configuration
The Stash platform requires an implementation of the Stash::Repo API for identifier assignment and submission to repositories.
Dryad uses CDL's EZID service for identifier assignment and stores datasets in the Merritt repository.
The Stash::Repo implementation is provided by the stash-merritt gem, which is included in the application Gemfile
and declared by the
repository: key in
EZID and Merritt/SWORD must be configured for each tenant in the apporpriate
tenants/*.yml file, e.g.
repository: # change me: you'll probably have to change all the following indented values and only if using Merritt repo type: merritt domain: merritt-repo-dev-example.cdlib.org endpoint: "http://uc3-mrtsword-dev.cdlib.org:39001/mrtsword/collection/my_collection_id" username: "submitter_username" password: "submitter_password" identifier_service: # change me: the identifier service is EZID here, may need to change this shoulder: "doi:10.5072/FK2" account: my_account_name password: my_account_password id_scheme: doi owner: null