GLEIF Dataset Librarian Integration Design

Add a new nullable column to ores_dq_artefact_types_tbl:

For datasets requiring no parameters, this column is NULL. For lei_parties, it contains:

{
  "type": "object",
  "properties": {
    "root_lei": {
      "type": "string",
      "description": "LEI code of the root entity to import",
      "minLength": 20,
      "maxLength": 20
    }
  },
  "required": ["root_lei"]
}

Add new artefact type rows for GLEIF publication:

The existing lei_entities and lei_relationships artefact types remain as staging-only (NULL target table and populate function).

All populate functions gain the p_params parameter:

CREATE OR REPLACE FUNCTION ores_dq_countries_publish_fn(
    p_tenant_id uuid,
    p_dataset_bundle_id uuid,
    p_params jsonb DEFAULT '{}'::jsonb
) RETURNS void AS $$ ... $$ LANGUAGE plpgsql;

Existing functions ignore p_params. The GLEIF party function extracts root_lei from it.

Straightforward bulk insert from dq_lei_entities_artefact_tbl to refdata_counterparties_tbl:

FOR each lei_entity IN artefact table:
    INSERT INTO refdata_counterparties_tbl (
        id, tenant_id, short_code, full_name, status, ...
    ) VALUES (
        gen_random_uuid(), p_tenant_id,
        lei_entity.lei,           -- short_code = LEI
        lei_entity.entity_legal_name,
        'active', ...
    )

No parameters required. Counterparty identifiers (LEI codes) are inserted into refdata_counterparty_identifiers_tbl as a separate step within the same function.

Multi-step function using the root_lei parameter:

STEP 1: Resolve subtree via recursive CTE on lei_relationships
    WITH RECURSIVE subtree AS (
        SELECT lei FROM dq_lei_entities_artefact_tbl WHERE lei = root_lei
        UNION ALL
        SELECT r.child_lei
        FROM dq_lei_relationships_artefact_tbl r
        JOIN subtree s ON s.lei = r.parent_lei
    )

STEP 2: Generate UUID map
    CREATE TEMP TABLE lei_uuid_map (lei TEXT, party_uuid UUID)
    INSERT INTO lei_uuid_map
        SELECT lei, gen_random_uuid() FROM subtree

STEP 3: Insert parties with parent resolution
    FOR each entity in subtree:
        parent_uuid = (SELECT party_uuid FROM lei_uuid_map
                       WHERE lei = relationship.parent_lei)
        INSERT INTO refdata_parties_tbl (
            id, tenant_id, short_code, full_name,
            parent_party_id, status, ...
        ) VALUES (
            lei_uuid_map[entity.lei], p_tenant_id,
            entity.lei, entity.entity_legal_name,
            parent_uuid,   -- NULL for root
            'active', ...
        )

STEP 4: Insert party identifiers
    FOR each entity in subtree:
        INSERT INTO refdata_party_identifiers_tbl (
            party_id, identifier_scheme, identifier_value, ...
        ) VALUES (
            lei_uuid_map[entity.lei], 'LEI', entity.lei, ...
        )

The root entity has parent_party_id = NULL, making it the tenant's root party (enforced by the existing partial unique index).

Modify the pipeline to pass parameters through to populate functions:

1. After resolving the topological sort order, check each dataset's artefact type for a non-NULL publication_params_schema.
2. Look up user-supplied parameters from the publication request (keyed by artefact type code).
3. Pass parameters as the p_params argument to the SQL populate function call.

The existing publication_result struct gains an optional params field per dataset. The SQL call changes from:

SELECT ores_dq_<type>_publish_fn(p_tenant_id, p_bundle_id);

to:

SELECT ores_dq_<type>_publish_fn(p_tenant_id, p_bundle_id, p_params);

• If lei_parties is invoked without root_lei, the function raises an exception with a clear message.
• If the root LEI does not exist in the artefact table, the function raises an exception.
• If the tenant already has a root party, the function raises an exception (the existing partial unique index enforces this at the DB level).

The "Configure LEI Parties" step needs a searchable list/table showing LEI entities from the artefact table. Key fields:

• LEI code
• Entity legal name
• Jurisdiction
• Entity status

The user selects one entity as the root. A search box filters by name or LEI code.

1. User selects "GLEIF-Based Tenant" option.
2. LEI entity picker appears (same component as the publish wizard).
3. User selects the root LEI entity.
4. Tenant name field auto-fills with entity_legal_name.
5. User can edit the name and description before confirming.
6. Provisioner creates the tenant, then publishes the base system bundle with root_lei set to the selected entity.