Schema

The structured story of the PDTF data model. Every leaf in the schema lives on exactly one aggregate page here — grouped by lifecycle and root authority, not by form or topic. Each leaf carries a declaration / evidence / derivation badge and the overlay chips that show every form-categorisation it participates in. Where the data dictionary is a flat browser and the business glossary is a concept index, this section is the narrative.

By the numbers

11 pages (1 section overview · 9 aggregate pages · 1 cross-cuts page)
3,561 leaves walked from the base transaction schema
21 diagrams across 6 types (ER · sequence · flowchart · state · gantt · pie)
15 overlays cross-referenced (3,463 leaf-to-overlay membership records)
2 canonical example transactions (London leasehold flat · Manchester freehold semi) — both validated against pdtf-transaction.json on every build

Start here

Overview

Section overview

The top-level entity map, the three kinds of fact (declaration / evidence / derivation), persona on-ramps, and the bindings between JSON Schema, OWL, SKOS, SHACL.

Process

Transaction & participants

The state machine and the people. Identity, role, capacity, address. ER diagram of participant attributes.

Chain, milestones & contracts

Process artefacts at the transaction level. ER + state diagram + Gantt timeline.

The property

Property

The implicit entity the schema never names. Four identity surfaces (address, UPRN, INSPIRE, title-linked) — documents a known structural defect.

Legal estate & title

HMLR-rooted aggregate: tenure, title number, proprietors, leases, charges, restrictions, notices. ER + sequence diagram for OC1 retrieval.

Property pack content

Built form, condition & valuation

The biggest aggregate. Built form, condition (cladding), fixtures & fittings, surveys, valuation, price. ER + sequence for survey-to-valuation feedback loop.

Utilities & energy

Heating, electricity, water, connectivity, EPC. Coupled by meters and fuel enums. ER + sequence for EPC issuance + pie for EPC band distribution.

Local context & searches

CON29R, CON29DW, LLC1, environmental, planning, neighbourhood. Sealed bundles from external authorities. Flowchart of request → authority → result + sequence.

Encumbrances & completion

Council tax, insurance, guarantees, occupiers, letting, completion statements. ER + apportionment flowchart + completion-day Gantt.

Cross-cutting

Evidence, documents & declarations

The claim layer. Documents, attachments, VC envelope, declaration metadata. ER + VC issuance sequence + three-kinds-of-fact flowchart.

Overlays, tasks & cross-cuts

Overlay leaf-counts for all 15 overlays, 6 task-shaped questions (cladding, knotweed, leasehold extension, probate, new-build, shared ownership), and the composition diagram showing how aggregates, overlays and tasks compose.

Mapping

Mapping to the ontology

How every leaf traces to its OWL term and back — the RML mapping, independently verified, 99.4% complete.

How it's built

Pages are generated by scripts/build-schema-pages.py: walk the base schema → classify each leaf via provenance-map.yaml → assign to canonical home via theme-map.yaml → bind canonical examples by JSON path → render Jinja2 templates with sidecar markdown narrative → drift-check 12 CI invariants. The pipeline is deterministic and re-runnable; the source documents are schema-section-ia.md and schema-content-generation-spec.md, both synthesised by 6-agent design swarms.