EA Data Model

The EA catalog exists in two forms: CSV (working/validation surface) and Cloudflare D1 (published consumption layer). CSVs are the editing surface — D1 is written only after the validation gate passes. The schemas are shared.

Objects CSV — `{slug}-objects.csv`

One row per architecture object. Schema version: v3.2 (current).

Column	Req	Values / Format	Description
ID	✓	{PREFIX}-NNN	Unique ID — prefix must match Type (e.g. APP-001, CAP-003)
Type	✓	See 18 prefixes	SAP LeanIX fact sheet type
Name	✓	Universal product name	Vendor/technical name as found in source (e.g. "Azure API Management")
Display_Name	✓	Organizational alias	What people in the org actually call it (e.g. "Maestro"). Never contains Name in parentheses.
Group	–	Architectural concept	Strategic category for board-level diagrams. Independent of vendor and org alias.
Subgroup	–	Sub-category	Optional further classification within a Group. No Group duplication.
Description	✓	Free text	What this component does
Level	✓	L0, L1, L2, L3	Hierarchy depth (L0 = top-level)
Parent_ID	–	ID of parent object	Hierarchy containment — enables visual nesting in diagrams
TOGAF_Layer	✓	Business, Application, Data, Technology	Architecture layer — drives diagram layer placement
Lifecycle	✓	plan, phaseIn, active, phaseOut, endOfLife	Current lifecycle status (LeanIX camelCase)
Business_Criticality	✓	missionCritical, businessCritical, businessOperational, administrativeService	Impact if unavailable
Functional_Fit	–	perfect, appropriate, insufficient, unreasonable	Functional suitability
Technical_Fit	–	fullyAppropriate, adequate, unreasonable, inappropriate	Technical suitability
Owner	–	Person or role name	Who owns this object
Department	–	Department name	Owning department
Tags	–	Comma-separated	Labels for filtering — include Macroscope A-code (e.g. A250, A370)
Cost_Size	–	XS, S, M, L, XL	Relative sizing
Cost_Amount	–	Number	Estimated cost
Notes	–	Free text	Additional context

Relations CSV — `{slug}-relations.csv`

One row per directed relationship between two objects.

Column	Req	Description
Source_ID	✓	ID of the source object (must exist in objects CSV)
Source_Display_Name	✓	Display name of the source object
Relationship_Type	✓	One of 12 standard types (contains, supports, requires, uses, realized-by, owned-by, provided-by, transfers, successor-of, data-flow, depends-on, triggers)
Target_ID	✓	ID of the target object (must exist in objects CSV)
Target_Display_Name	✓	Display name of the target object
Description	✓	What this relationship means in context
Visual_Verb	✓	Short label for diagram edges (e.g. "uses", "triggers", "data flow")

CO CSV — `{slug}-co.csv`

Changement Organisationnel — optional, produced only for roadmap-driven transformations. Mandatory for migration projects (one CO row per decommissioned, introduced, or migrated object).

Column	Req	Values / Description
CO_ID	✓	CO-NNN — unique change record ID
Object_ID	✓	ID of the object being changed
Initiative_ID	✓	ID of the Initiative driving this change (INI-NNN) — use program-level INI
Scenario	–	Option_A, Option_B… — for orientation comparisons. Omit for single-project changes.
Impact_Type	✓	introduction, decommissioning, rollout, withdrawal, scopeChange, technicalChange, interfaceChange, discontinuation
Change_Type	✓	New, Modified, Retired, Migrated, Consolidated
Impact_Level	✓	High, Medium, Low
Impact_Description	✓	What changes and why
Current_State	✓	As-is state description
Target_State	✓	To-be state description
Phase	–	Implementation phase (e.g. "Phase 0 — Foundation" or "INI-002")
Cost_Project	–	Project cost estimate
OpEx_Delta	–	Change in operational cost
CapEx_Delta	–	Change in capital cost
Status	✓	Planned, In_Progress, Completed, Cancelled
Notes	–	Additional context

D1 Database Structure

One D1 database per customer. Each database holds all that customer's DAE catalogs via a catalog discriminator column. The same SQLite schema runs locally and on Cloudflare (foundry split).

catalogs

Registry of DAE catalogs per customer

id, dae, slug, label, customer

objects

All EA objects across catalogs

id, catalog, type, name, display_name, group, togaf_layer, lifecycle, attrs (JSON)

relations

All relationships between objects

id, catalog, source_id, target_id, type, description, visual_verb

-- Key queries (Pages Function API)

GET /api/catalogs

→ all catalogs for the current customer

GET /api/objects?catalog=TES

→ all objects in catalog TES (Transgesco EA System)

GET /api/relations?catalog=TES

→ all relations in catalog TES

GET /api/orphans

→ objects with no relations and no Parent_ID (should return empty)

GET /api/stats

→ counts per type, lifecycle, TOGAF layer

Customer Config — `customers.json`

Located at experiments/dae-catalog-d1/customers.json. Adding a customer or DAE is a config edit — no code change needed.

{
  "transgesco": {
    "db": "transgesco",
    "daes": [
      { "dae": "DAE-0007", "catalog": "TES", "label": "Évaluation des systèmes" },
      { "dae": "DAE-0007", "catalog": "BCM", "label": "Business Change Map" }
    ]
  },
  "stm": {
    "db": "stm",
    "daes": [
      { "dae": "DAE-0001", "catalog": "VTB", "label": "Vigie technique BizTalk" },
      { "dae": "DAE-0002", "catalog": "PIS", "label": "Passeport intégration STM" }
    ]
  }
}

Schema Version History

v3.2 Current Mandatory INI sweep for all AS-IS extractions + Macroscope deliverable routing (A-codes in Tags)

v3.1 Migration Project Mode + orchestration granularity rules (PRO/PRS split) + PRS as 18th type

v3.0 Full LeanIX v4 coverage: 17→18 prefixes, CO Impact_Type + Scenario, camelCase lifecycle values

v2.1 Display_Name (renamed), Group + Subgroup columns, Functional_Fit added

v2.0 English columns, Level, Parent_ID, Lifecycle, Business_Criticality added

v1.0 Legacy French columns (Nom, Couche TOGAF…) — auto-migrated on extraction

EA Data Model

Objects CSV — {slug}-objects.csv

Relations CSV — {slug}-relations.csv