Skip to content

OpenSKOS/OpenSKOS

Repository files navigation

Build Status

1. Install the OpenSKOS code


Docker development

Docker images were added to the OpenSkos project in 2017. If you wish to develop using docker, please proceed to section 8


Copy the code to a location of your choice.

Make sure all files are readable by your webserver. Make sure the directories data/uploads, cache, public/data/icons/assigned and public/data/icons/uploads are writable for the webserver.

For security reasons you can place the data directory outside your webserver's document root.

1.1 Composer

Run composer install to install some dependencies such as zend framework 1.12

1.2 Configuration

To configure OpenSKOS you have to rename:

APPROOT/application/configs/application.ini.dist

to

APPROOT/application/configs/application.ini

Now you can edit the APPROOT/application/configs/application.ini.

You can have separate config settings for specific deployments. The configuration section marked by the Environment Variable APPLICATION_ENV (see 2.1 Setting Up Your VHOST). Most settings are self explanatory.

If you experience any problems you may want to modify settings in the config, to show you more verbose error messages:

    resources.frontController.params.displayExceptions=1
    phpSettings.display_errors = 1

1.2.1 OAI-PMH setup

OpenSKOS includes a OAI harvester. To configure OAI Service providers, use the "instances" part of the configuration. Two types of instances are supported:

  • openskos (instances of OpenSKOS)
  • external (any OAI-PMH provider that provides SKOS/XML-RDF data)

The setup for "openskos" types is easy:

    instances.openskos.type=openskos
    instances.openskos.url=http://HOSTNAME
    instances.openskos.label=YOUR LABEL

For "external" types use this syntax:

    instances.example1.type=external
    instances.example1.url=http://HOSTNAME
    instances.example1.label=EXAMPLE LABEL
    #optional, default=oai_rdf
    instances.example1.metadataPrefix=METADATAPREFIX
    #optional:
    instances.example1.set=SETSPEC

You can define multiple instances by using a different key (in the above example the key example1 is used).

1.2.2 ConceptScheme ordering

The application.ini allows you to change the order in which concept schemes are listed everywhere. The scheme order is made in this sequence:

  • group the schemes according to their collection
  • order the groups by the desired collection order
  • sort the schemes inside each group alphabetically

The collection order can be set in the ini by setting the editor.schemeOrder.collections[]="<collectionUri>" All unlisted collections will be ordered after the listed ones. All listed collections that re not present in the DB will be skipped. In this way the ini supports collection ordering for more than 1 instances.

2. Webserver with PHP support

You can install your favourite webserver with PHP support. All development and testing was done using Apache/2.2.15 with PHP 5.3.8 Make sure your PHP installation supports at least one supported Database adapters (see http://framework.zend.com/manual/en/zend.db.adapter.html or otherwise: https://docs.zendframework.com/zend-db/adapter/)

2.1 Setting Up Your VHOST

The following is a sample VHOST you might want to consider for your project.

<VirtualHost *:80>
   DocumentRoot "/PATH/TO/CODE/public"
   ServerName YOUR.SERVER.NAME

   # This should be omitted in the production environment
   SetEnv APPLICATION_ENV development

   <Directory "/PATH/TO/CODE/public">
       Options Indexes MultiViews FollowSymLinks
       AllowOverride All
       Order allow,deny
       Allow from all
   </Directory>

</VirtualHost>

3. Database setup

Install your choice of Zend Framework supported Database engine (see http://framework.zend.com/manual/en/zend.db.adapter.html or otherwise: https://docs.zendframework.com/zend-db/adapter/). The credentials to access your database can be configured in the application's configuration.

Once you have created an empty database, you have to run the SQL script APPROOT/data/openskos-create.sql to create the db-tables.

You also have to run the php-script to create a tenant:

php APPROOT/tools/tenant.php --code INST_CODE --name INST_NAME --email EMAIL --password PWD create

With this account created you can login into the dashboard, where you can manage all the other entities of the application.

4. Apache Jena Fuseki setup

Openskos is compatible with Fuseki 2 or Fuzeki 3 for storage. It has been tested up to Fuzeki 3.8 (the latest stable version at time of writing)

Installing Fuseki 2 for development purposes:

  1. Download Fuseki 2 from here download
  2. Install stand alone fuseki server. The instructions are the same as on getting started page
  3. Unpack the downloaded file with unzip or tar zxfv to a <fuseki folder> of your choice
  4. chmod +x fuseki-server
  5. Symlink or copy the content of: <openskos folder>/data/fuseki/configuration/ to <fuseki folder>/run/configuration/
  6. Go to <fuseki folder> and run the server with ./fuseki-server --update
  7. The docs say that Fuseki requires Java 7, but if you have the error Unsupported major.minor version 52.0 try updating your Java, or go for Java 8 directly.
  8. Now you will have the fuseki server up and running on http://localhost:3030/ with "openskos" dataset defined. This is also the default config in openskos' application.ini.dist - item sparql

4.1 Jena Updates

Several bug fixes were made to the rules/openskos.ttl file in October 2018, on both the OpenSkos 2.2 (Master at time of update ) and Meertens Merge (Development at time of update) branches. When upgrading to these versions, please update the configuration files on the Jena server to the versions located in ./data/fuseki/configuration.

5. Apache Solr Setup

You have to have a java VM installed prior to installing Solr! The version of Solr used during development was 7.4.0. Other versions going back to Solr 4 are supported, but it will be necessary to adapt the Solr configuration files to the syntax for these versions.

http://www.apache.org/dyn/closer.cgi/lucene/solr/

  • go to the example directory and create a directory named openskos
  • copy the data/solr/conf directory of the OpenSKOS checkout to the SOLR-INSTALL_DIR/example/openskos directory

You can now start Solr (in this example with 1,024 MB memory assigned):

    java -Dsolr.solr.home="./openskos" -Xms1024m -Xmx1024m -jar start.jar

5.1 Solr Updates

The Solr configuration file was substantially altered during the Meertens Merge project. If upgrading, you will need to take the example solrconfig.xml and schema.xml from ./data/solr, and adapt them to your Solr version.

After updating the configuration, you should delete the contents of the Solr database and re-index using the ./tools/jena2Solr.php script. If you skip this step, OpenSkos will remain functional, but the internal content of the Solr core will become inconsistent as records are updated.

6. Data Ingest

Once you have the application running you can start adding data, managed in collections.

You can create a collection in the dashboard.

There are three ways to populate a collection:

6.1 REST-interface

Send data via the REST-API, e.g. like this:

curl -H "Accept: text/xml" -X POST -T sample-concept.rdf http://localhost/OpenSKOS/public/api/concept

You find the required format of the input data described in the API-docs under: http://openskos.org/api#concept-create

You may send only one concept per call. Also, you have to identify the tenant and provide the API key, which you assign to the user in the dashboard.

6.2 Uploader

Upload a dataset (a SKOS/RDF file) via a form in the dashboard:Manage collections. Here you can provide many concepts within one file (XPath: /rdf:RDF/rdf:Description)

Once you successfully upload the file, it is scheduled for import, as seen in dashboard:Manage jobs.

The import job can be started with ./tools/jobs.php, a CLI script intended to be run with a Cron like task runner.

6.3 OAI ???

Third possiblity is to replicate an existing dataset via OAI-PMH, either from other OpenSKOS-instances or from an external source providing SKOS-data.

??? For this, you set the [OAI baseURL]-field of a collection to the OAI-PMH endpoint of an external provider and let the source be harvested.

The harvest job can be started with ./tools/harvest.php, another CLI script meant to be run as a cron-task. ???

6.4 Migrate from OpenSKOS v1

It is possible to migrate the data from the SOLR core used by a OpenSKOS v1 instance directly into a v2 instance

tools/migrate.php --endpoint http://<solr server>:8180/ciss/<core name>/select

Once this is complete the data from the v1 instance will be available in the triple store used by OpenSKOS v2.

6.5 API Documentation

Generate API Documentation

npm install
npm run doc

Visit: http://example.com/apidoc/

6.6. Using the API

Full HTML documentation of the API is supplied and is available in HTML at <baseruri>/apidoc

7. Development

7.1. Migration from OpenSKOS-1 to OpenSKOS-2.2

WARNING: It is very strongly recommended to back up all data before performing the following steps

In OpenSkos 2.2 Tenants and Collections in MySQL have been migrated from MySQL to the Jena triple store.

To migrate from OpenSKOS 1.0 or 2.1 to 2.2, first read sections 4.1 and 5.1 about updating the Jena and Solr configurations. Both steps are necessary when upgrading to OpenSkos 2.2

Then perform the following steps:

-- /tools/migrate_tenant_collection.php (migrates tenants and collections from MySQL to institutions and sets of Triple store)

-- optionally /tools/labelsToXl.php (this is a picturae script slightly extended by Meertens), if skos xl labels are demanded.

Examples of the corresponding command lines are:

php migrate_tenant_collection.php --db-hostname=localhost --db-database=geheim 
--db-password=geheim --db-username=ookgeheim --debug=1

Adding skos xl labels is also possible since version 2.1. To activate, first edit the tenant to enable SkosXL, and then update Jena with:

labelsToXl.php –add=1

The SOLR schema.xml file has been updated in version 2.2. Having completed the migration, please empty the core, and then update the schema.xml file. Then fill the Solr database with the script:

php tools/jena2solr.php

Max notations are now maintained in a separate MySQL table. Use the script ./data/dbchanges/20180724.sql to update the db schema.

Then execute: php tools/updateMaxNotation.php

-- Publisher URI And extra triple needs to be added to concepts to allow continued functioning of OpenSKOS.

php tools/fillConceptPublisher.php

7.2. Import (/tools/skos2openskos.php)

Example of a command line:

php skos2openskos.php --setUri=http://htdl/clavas-org/set 
--userUri=http://localhost:89/clavas/public/api/users/4d1140e5-f5ff-45da-b8de-3d8a2c28415f 
--file=clavas-organisations.xml

8 Development using Docker

TL;DR;

composer install [--ignore-platform-reqs]
php vendor/bin/phing config

docker network create openskos
docker-compose up 

# go to localhost:9001 and create a dataset matching in name with application.ini sparql.queryUri if it's missing

sudo chmod 777 data/solr
sudo chmod 666 data/solr/*
sudo chown -R 8983:8983 data/solr
docker exec -it openskos-php-fpm php vendor/bin/phing solr.create.core
docker exec -it openskos-php-fpm php tools/tenant.php --code CODE --name NAME --email EMAIL --password PASSWORD create
docker exec -it openskos-php-fpm php tools/jena2solr.php

# go to localhost:9000 and log in using your just-created credentials

8.1 Installing docker

To test / develop the application go to the root folder, and run:

docker-compose up
composer install
docker exec -it openskos-php-fpm ./vendor/bin/phing install.dev

8.2 Updating the configuration

Then copy the file ./application/configs/application.ini.dist to ./application/configs/application.ini

Under the section ; Solr configuration: add the following:

solr.host=openskos-solr
solr.port=8983
solr.context=/solr/openskos

Under the section '; Apache Jena Fuseki configuration:' add the following

sparql.queryUri=http://openskos-jena-fuseki:3030/openskos/query
sparql.updateUri=http://openskos-jena-fuseki:3030/openskos/update

Configuring Jena

You can then create an empty graph with the name openskos in the Jena interface at http://localhost:9001. The admin username:password combination is admin and admin. Here you can create a persistent dataset named openskos

Under the section ; Database configuration: add the following

resources.db.adapter=pdo_mysql
resources.db.params.host=openskos-mysql
resources.db.params.username=root
resources.db.params.password=secr3t
resources.db.params.charset=utf8
resources.db.params.dbname=openskos
resources.db.isDefaultTableAdapter = true

8.3 Running OpenSkos

Create a test tenant / user in the openskos application

docker exec -it openskos-php-fpm php ./tools/tenant.php create -e development --code=pic --name=Picturae --email=test@example.com --password=test

Now you can login on http://localhost:9000/editor/login