~2.9, ~2.10, ~2.11

Install via pip:

pip install ckanext-create-typed-package

Enable the plugin:

ckan.plugins = create_typed_package

Note: Extension requires pre-configured custom dataset types (via ckanext-scheming or CKAN type registry).

Example with ckanext-scheming:

pip install ckanext-scheming

ckan.plugins = scheming_datasets create_typed_package

scheming.dataset_schemas = path/to/schema1.yaml path/to/schema2.yaml

Configuration options:

Build type list using ckanext-scheming API instead of CKAN’s package_type registry (optional, default: false)

create_typed_package.use_scheming = true

Additional types not automatically added to the list (optional, default: [])

create_typed_package.additional_types = custom_type another_type

Exclude specific types from available list (optional, default: [])

create_typed_package.exclude_types = unwanted_type another_unwanted

Use separate page instead of modal for type selection (optional, default: false)

create_typed_package.use_separate_route = true

Replace ‘Add Dataset’ button with link to type selection (requires use_separate_route=true, optional, default: false)

create_typed_package.replace_dataset_button = true

Custom URL for type selector page (optional, default: /dataset/select-type)

create_typed_package.route_path = /create-package/select-type

Custom labels for dataset types (optional, default: tk._(type_machine_name))

create_typed_package.label_for. = Human Readable Label

create_typed_package.label_for.dataset = Publication

Dynamic type options (advanced):

Implement IAction interface and register chained action ‘ctp_list_types’ to provide context-dependent type lists:

import ckan.plugins.toolkit as tk
from ckan.types import Context

@tk.chained_action
@tk.side_effect_free
def ctp_list_types(next_action, context: Context, data_dict):
    if tk.get_endpoint() == ("organization", "read"):
        organization_id = tk.request.view_args.id
        allowed_types = ...  # compute types for organization
        return [
            {"name": item.VALUE, "label": item.LABEL}
            for item in allowed_types
        ]
    return next_action(context, data_dict)

Action must return list of dicts with ‘name’ (machine name) and ‘label’ (human readable).

Behavior: - If action returns >1 item: modal/page with selector shown - If action returns exactly 1 item: redirect directly to that type’s form

UI modes: 1. Modal (default): JavaScript-based modal after clicking ‘Add Dataset’ 2. Separate page: /dataset/select-type (enable via use_separate_route) 3. Button replacement: Replace original button with selector link (requires use_separate_route)

Requirements: - Python >= 3.8 - Custom dataset types configured (ckanext-scheming recommended) - ckantoolkit dependency