A CKAN extension to generate the /data.json file and to harvest data sources from a remote /data.json file according to the U.S. Project Open Data metadata specification (http://project-open-data.github.io/).
This plugin creates a new view at /data.json (or other configurable path) that outputs the contents of the data catalog in the Project Open Data JSON metadata format. It also creates a view at /data.jsonld which outputs the same in JSON-LD format.
The plugin also provides a harvester to import datasets from other remote /data.json files. See below for setup instructions.
And the plugin also provides a new view to validate /data.json files at http://ckanhostname/pod/validate.
This module assumes metadata is stored in CKAN in the way we do it on http://hub.healthdata.gov. If you're storing metadata under different key names, you'll have to revise ckanext/datajson/plugin.py accordingly.
To install, activate your CKAN virtualenv, install dependencies, and install the module in develop mode, which just puts the directory in your Python path.
. path/to/pyenv/bin/activate
pip install -r pip-requirements.txt
python setup.py develop
Then in your CKAN .ini file, add ``datajson'' to your ckan.plugins line:
ckan.plugins = (other plugins here...) datajson
That's the plugin for /data.json output. To make the harvester available, also add:
ckan.plugins = (other plugins here...) harvest datajson_harvest
If you're running CKAN via WSGI, we found a strange Python dependency bug. It might only affect development environments. The fix was to revise wsgi.py and add:
import ckanext
before
from paste.deploy import loadapp
Then restart your server and check out:
http://yourdomain.com/data.json
and
http://yourdomain.com/data.jsonld
and
http://yourdomain.com/pod/validate
If you're deploying inside Apache, some caching would be a good idea because generating the /data.json file can take a good few moments. Enable the cache modules:
a2enmod cache
a2enmod disk_cache
And then in your Apache configuration add:
CacheEnable disk /data.json
CacheRoot /tmp/apache_cache
CacheDefaultExpire 120
CacheMaxFileSize 50000000
CacheIgnoreCacheControl On
CacheIgnoreNoLastMod On
CacheStoreNoStore On
And be sure to create /tmp/apache_cache and make it writable by the Apache process.
Generating this file is a little slow, so an alternative instead of caching is to generate the file periodically (e.g. in a cron job). In that case, you'll want to change the path that CKAN generates the file at to something other than /data.json. In your CKAN .ini file, in the app:main section, add:
ckanext.datajson.path = /internal/data.json
Now create a crontab file ("mycrontab") to download this URL to a file on disk every ten minutes:
0-59/10 * * * * wget -qO /path/to/static/data.json http://localhost/internal/data.json
And activate your crontab like so:
crontab mycrontab
In Apache, we'll want to block outside access to the "internal" URL, and also map the URL /data.json to the static file. In your httpd.conf, add:
Alias /data.json /path/to/static/data.json
<Location /internal/>
Order deny,allow
Allow from 127.0.0.1
Deny from all
</Location>
And then restart Apache. Wait for the cron job to run once, then check if /data.json loads (and it should be fast!). Also double check that http://yourdomain.com/internal/data.json gives a 403 forbidden error when accessed from some other location.
You can customize the URL that generates the data.json output:
ckanext.datajson.path = /data.json
ckanext.datajsonld.path = /data.jsonld
ckanext.datajsonld.id = http://www.youragency.gov/data.json
If ckanext.datajsonld.path is omitted, it defaults to replacing ".json" in your ckanext.datajson.path path with ".jsonld", so it probably won't need to be specified.
The option ckanext.datajsonld.id is the @id value used to identify the data catalog itself. If not given, it defaults to ckan.site_url.
You can also set some default values for datasets that are missing required fields. These are possible:
ckanext.datajson.default_contactpoint = Health Data Initiative
ckanext.datajson.default_mbox = [email protected]
ckanext.datajson.default_keywords = health
To use the data.json harvester, you'll also need to set up the CKAN harvester extension. See the CKAN harvester README at https://github.com/okfn/ckanext-harvest for how to do that. You'll set some configuration variables and then initialize the CKAN harvester plugin using:
paster --plugin=ckanext-harvest harvester initdb --config=/path/to/ckan.ini
Now you can set up a new DataJson harvester by visiting:
http://yourdomain.com/harvest
And when configuring the data source, just choose "/data.json" as the source type.
The next paragraph assumes you're using my fork of the CKAN harvest extension at https://github.com/JoshData/ckanext-harvest
In the configuration field, you can put a YAML string containing defaults for fields that may not be set in the source data.json files, e.g. enter something like this:
defaults:
Agency: Department of Health & Human Services
author: Substance Abuse & Mental Health Services Administration
author_id: http://healthdata.gov/id/agency/samhsa
The keys title, notes, author, and url are stored in the corresponding CKAN
package fields. tags, which must be an array of strings, is stored as CKAN package
tags. All other keys are stored in package extras (i.e. the Additional Info of a
package).
You may also override values found in the harvested file with fixed values set in the configuration like so:
overrides:
author: U.S. Food and Drug Administration
You can also specify filters to control what datasets from the data.json file are imported. By default everything is imported. In a "filters" section, map data.json field names to an array of permitted values or in an "excludes" section map data.json field names to values or regular expressions surrounded in forward slashes that will cause datasets to be excluded:
filters:
theme: ["Health"]
excludes:
accessURL: ["http://example.org/dataset", "/^http://some-pattern/here/"]
Each dataset is compared to each filter and exclude (in this example, a theme filter and an accessURL exclude). To be imported, the dataset must match at least one value of a filter and may not match any value of an exclude. In this example, all imported datasets will have Health set as their theme and no accessURL will be either "http://example.org/dataset" or start with "http://some-pattern/here" (not that the final slash indicates the end of the regular expression).
Written by the HealthData.gov team.
As a work of the United States Government, this package is in the public domain within the United States. Additionally, we waive copyright and related rights in the work worldwide through the CC0 1.0 Universal public domain dedication (which can be found at http://creativecommons.org/publicdomain/zero/1.0/).