ckanext-syndicate

CKAN plugin to syndicate datasets to another CKAN instance.

This plugin provides a mechanism to syndicate datasets to another instance of CKAN. If a dataset has the syndicate flag set to True in its custom metadata, any updates to the dataset will be reflected in the syndicated version. Resources in the syndicated dataset are stored as the URLs of the resources in the original. You must have the API key of a user on the target instance of CKAN. See the Config Settings section below.

Plugins can modify data sent for syndication or react to before/after syndication events by implementing the ISyndicate interface and subscibing to corresponding signals. This is useful if the schemas are different between CKAN instances.

Warning

Starting from v2.0 ckanext-syndicate supports only CKAN greater or equal v2.9 and python versions starting from v3.7. In addition, it was reworked and does not relies on actions as it was before. If you need old behavior or python2 support, switch to any of v1 tags (v1.1.0).

Requirements

• Tested with CKAN 2.9.x branch on python v3.7+
• To work over SSL, requires pyOpenSSL (pip install -r requirements.txt)

Installation

To install ckanext-syndicate:

1. Activate your CKAN virtual environment, for example::

    . /usr/lib/ckan/default/bin/activate
   

2. Install the ckanext-syndicate Python package into your virtual environment::

    git clone https://github.com/aptivate/ckanext-syndicate.git
    pip install -e ckanext-syndicate
   

3. Add syndicate to the ckan.plugins setting in your CKAN config file.

Config Settings for using in .ini file

Syndication perfomrs dataset creation and update on the remote portal. It also possible to syndicate the dataset to the multiple portals simultaneously. ckanext-syndicate makes no assumptions as to how many syndication endpoints you have and performs each synchronization separately as if you've configured the first syndication endpoind, did a syndication, updated configuration, did a syndication once again.

Internally, set of config option related to the particular endpoint is called profile(ckanext.syndicate.types.Profile). Each profile has an ID. ID is a part of config option: ckanext.syndicate.profile.<PROFILE ID>.<OPTION> If you want to syndicate dataset to the two different portals, first and another, configuration may look like:

ckanext.syndicate.profile.first.ckan_url = https://data.example.com
ckanext.syndicate.profile.another.ckan_url = https://another.example.com

Here is the full list of config options available for Profile. Dont forget to replace PROFILE_ID with the any identifier you like.

 # The URL of the site to be syndicated to
 # (required)
 ckanext.syndicate.profile.PROFILE_ID.ckan_url = https://data.example.com

 # The API key of the user on the syndicated site
 # (required)
 ckanext.syndicate.profile.PROFILE_ID.api_key = 9efdd954-c643-444a-97a1-c9c374cef861

 # The custom metadata flag used for syndication
 # (optional, default: syndicate).
 ckanext.syndicate.profile.PROFILE_ID.flag = syndicate_to_hdx

 # The custom metadata field to store the syndicated dataset ID
 # on the original dataset
 # (optional, default: syndicated_id)
 ckanext.syndicate.profile.PROFILE_ID.field_id = hdx_id

 # A prefix to apply to the name of the syndicated dataset
 # (optional, default: '')
 ckanext.syndicate.profile.PROFILE_ID.name_prefix = my-prefix

 # The name of the organization on the target CKAN to use when creating
 # the syndicated datasets
 # (optional, default: None)
 ckanext.syndicate.profile.PROFILE_ID.organization = my-org-name

 # Try to preserve dataset's organization
 # (optional, default: false)
 ckanext.syndicate.profile.PROFILE_ID.replicate_organization = yes

 # The username whose api_key is used.
 # If the dataset already exists on the target CKAN instance, the dataset will be updated
 # only if this option is set and its creator matches this user name
 # (optional, default: None)
 ckanext.syndicate.profile.PROFILE_ID.author = some_user_name

Extending

Syndication can be configured for the each individual portal. There are two types of customization: reactions on events and changes to workflow.

Reactions are usefull when you need to perform a side-effect right before or right after the syndication. This can be achieved via the blinker's signals. ckanext-syndicate provides two signals that can be imported from the ckanext.syndicate.signals (or subscribed via ISignal starting from CKAN v2.10):

• before_syndication
• after_syndication

Both signals get local dataset's ID as sender and extra keyword argument profile(current syndication profile). Basic subscription looks like this:

@after_syndication.connect
def after_syndication_listener(package_id, **kwargs):
	profile = kwargs.get("profile")
	if profile:
		do_something(package_id, profile)

Changes to syndication workflow are made via ckanext.syndicate.interfaces.ISyndicate interface. At moment, it contains two methods:

• skip_syndication - decide, whether syndication must be performed for the given profile.
• prepare_package_for_syndication - update the package, before it sent to the remote portal. It can be really usefull if portal that you are syndicating to, is using different metadata schema.

Basic implementations looks like this:

class MyPlugin(plugins.Plugin):
	plugins.implements(ISyndicate, inherit=True)

	def skip_syndication(
		self, package: model.Package, profile: Profile
	) -> bool:
		if should_be_syndicated(package):
			return False
		return True

	def prepare_package_for_syndication(
		self, package_id: str, data_dict: dict[str, Any], profile: Profile
	) -> dict[str, Any]:
		data_dict.pop("sensitive_field")
		return data_dict

Default implementation of skip_syndication prevents syndication for:

• private datasets
• datasets with the falsy value of the field, specified by ckanext.syndicate.profile.PROFILE_ID.flag config option(syndicate by default)

CLI

Mass or individual syndication can be triggered as well from command line::

ckan syndicate sync [ID]

Running the Tests

Install dev-requirements.txt:

pip install -r dev-requirements.txt

Run the tests:

pytest --test-ini ckan.ini