Assessment Schema YAML

Assessment Schema YAML

Guide for editing YAML assessment sources (sources/prescan.yaml, sources/dpia.yaml, sources/iama.yaml) that conform to schemas/assessment-definition.v1.schema.json.

Top-level Structure

name: "DPIA Rapportagemodel Rijksdienst"
urn: "urn:nl:dpia:3.0"           # Pattern: urn:nl:<type>:<version>
version: "3.0"
description: "..."
tasks: [...]                      # Required: array of task objects
assessments: [...]                # Optional: evaluation rules (used in prescan)
prefixQuestionIds: true           # Optional: prefix every label with its official question ID (used by iama.yaml)

The IAMA (sources/iama.yaml, urn: "urn:nl:iama", version: "2.0") is a third assessment type alongside Pre-scan and DPIA. It sets prefixQuestionIds: true, which prefixes each task label with its official question ID.

Task Structure

Every task requires these fields:

- task: "Task title"              # Required: display name
  id: "2.1.3"                    # Required: hierarchical ID, pattern ^[0-9]+(\.[0-9]+)*
  type:                           # Required: array of field types
    - open_text
  repeatable: false               # Required: boolean

Optional fields:

• description: explanation text (may contain HTML)
• category: grouping category
• is_official_id: true if ID matches the Rapportagemodel DPIA numbering
• valueType: expected data type (string, boolean, number, string[], boolean|null)
• defaultValue: pre-filled value
• options: required when type includes select_option, checkbox_option, or radio_option
• tasks: nested child tasks (for task_group type)
• dependencies: conditional display rules
• calculation: score calculation logic
• references: cross-references to other assessments
• sources: external resource references (images, documents)
• required_status: whether the field is mandatory
• instance_label_template: template for repeatable instance labels (e.g. "Gegevensverwerking {4.1.1}")

IAMA-specific task fields (present in sources/iama.yaml):

• in_fria: true — marks a question as part of the FRIA (Fundamental Rights Impact Assessment, art. 27 AI-verordening)
• action_point_group: true — marks a task group that collects the action points ("Actiepunten") for a deel
• action_point_summary: true — marks the deel (Deel 5 – Afsluiting) that summarises all action points

Field Types

Type	Description	Requires
text_input	Single-line text	-
open_text	Multi-line text	-
select_option	Dropdown select	options array
multiselect_scrollable	Scrollable multi-select list (used in iama.yaml)	options array
checkbox_option	Multiple checkboxes	options array
radio_option	Radio buttons	options array
task_group	Container for child tasks	tasks array
informational	Read-only informational block (no input; used in iama.yaml)	-
date	Date picker	-

A task can have multiple types (array), e.g. [open_text, select_option].

Options

options:
  - value: "option_value"         # Required: string, boolean, or null
    label: "Display label"        # Optional: if omitted, value is shown

Dependencies

Three types of dependencies control conditional behavior:

conditional — Show/hide based on another answer

dependencies:
  - type: conditional
    condition:
      id: "2.1.6"                # Task ID to check
      operator: equals            # equals | contains | any | all
      value: true                 # Value to match
    action: show                  # show | hide

source_options — Load options dynamically from another field

dependencies:
  - type: source_options
    condition:
      id: "2.1"                  # Source task ID
    action: options

instance_mapping — Sync instances between repeatables

dependencies:
  - type: instance_mapping
    source:
      id: "3.1"                  # Source repeatable task
    mapping_type: one_to_one      # one_to_one | one_to_many | many_to_one
    action: sync_instances

Calculations (Pre-scan)

Score calculations use jexl expressions:

calculation:
  scoreKey: "gewone_persoonsgegeven"
  expression: "answers('1.1.2') | count"
  riskScore:
    - when: "gewone_persoonsgegeven < 6"
      value: 0
    - when: "gewone_persoonsgegeven >= 6"
      value: 1

Available jexl functions

• answers('taskId') — get the answer value for a task
• bool(value) — convert to boolean
• count — filter/transform that counts array items
• countSelectedOptions('taskId') — count selected checkbox options
• weightedCountMap(answers, weightMap) — weighted count of selected options

References

Cross-references between Pre-scan and DPIA:

references:
  prescanModelId: "1"             # Links to Pre-scan Model paragraph
  DPIA: "1.1"                    # Simple: links to DPIA task ID
  DPIA:                           # With type:
    - id: "1.1"
      type: "one-to-one"         # one-to-one | one-to-many | many-to-one | many-to-many
  DPIA:                           # Typed reference:
    id: "12.1.1"
    type: "pre-fill"             # pre-fill | pre-view | direct takeover

Assessment Evaluation Rules (Pre-scan)

assessments:
  - id: "DPIA"
    levels:
      - level: "required"         # required | recommended
        expression: "criteria.wetgeving || criteria.riskscore"
        result: "DPIA verplicht"
        criteria:
          - id: "wetgeving"
            expression: "bool(answers('0.1'))"
            explanation: "er sprake is van nieuwe wet- of regelgeving"

ID Numbering

• Top-level sections: single number ("0", "1", "2")
• Subsections: dot-separated ("2.1", "2.1.3")
• Task ID "0" is the introduction (not an official DPIA section)
• is_official_id: true means the ID corresponds to the Rapportagemodel numbering

Validation

After editing YAML sources, validate with:

python script/schema_validator.py --schema schemas/assessment-definition.v1.schema.json --source sources/dpia.yaml
python script/schema_validator.py --schema schemas/assessment-definition.v1.schema.json --source sources/prescan.yaml
python script/schema_validator.py --schema schemas/assessment-definition.v1.schema.json --source sources/iama.yaml

The schema file schemas/assessment-definition.v1.schema.json is the source of truth. Always read it when unsure about allowed values.