Skip to main content
Jira progress: loading…

UOM-TYPE

UOM Engines

1. Overview​

UOM Engines are Micro Engines responsible for computing intelligence from the Universal Organizational Model.

They operate across ZAYAZ’s canonical organizational and physical graph:

Legal Entity
  → Facility
  → Building
  → Resource Point
  → Accepted Measurement

UOM engines transform Resource Points, relationships, facilities, buildings, leases, occupancy snapshots, and accepted measurements into reusable operational and ESG intelligence.

UOM is an engine type in the MICE taxonomy. Concrete UOM engines are implemented as MEID-registered micro engines.

2. Design Principles​

  1. Graph-Aware
    • UOM engines must understand entity relationships and Resource Point relationships.
  2. Accepted Values Only
    • UOM engines consume accepted measurements, not raw observations, unless explicitly designed for reconciliation or completeness checks.
  3. Tenant-Scoped
    • Every execution must be strictly scoped by client_id.
  4. No Duplicate Counting
    • UOM engines must use Resource Points as canonical measurement boundaries.
  5. Reusable by ZARA
    • Outputs must be structured, explainable, and suitable for natural-language responses.

3. Scope of Responsibility​

UOM engines perform computations over organizational, physical, and resource structures.

Supported responsibility areas include:

  • Resource aggregation
  • Intensity calculations
  • Occupancy resolution
  • Resource coverage analysis
  • Measurement completeness checks
  • Resource reconciliation
  • Facility-level intelligence
  • Building-level intelligence
  • Entity-level resource rollups

4. What UOM engines do not do​

UOM engines do not directly perform:

  • ❌ OCR
  • ❌ Raw invoice extraction
  • ❌ AI narrative generation
  • ❌ Framework interpretation
  • ❌ Final assurance approval
  • ❌ Unit normalization where MEID_CORE_NORM is available
  • ❌ Product carbon allocation unless part of PCF engines

5. Canonical UOM Engine Set​

Initial UOM engines:

  • MEID_UOM_001 Resource Aggregation Engine
  • MEID_UOM_002 Intensity Calculation Engine
  • MEID_UOM_003 Occupancy Engine
  • MEID_UOM_004 Resource Coverage Engine
  • MEID_UOM_005 Measurement Completeness Engine
  • MEID_UOM_006 Resource Reconciliation Engine

6. Inputs​

UOM engines consume:

  • UOM graph nodes
  • UOM graph edges
  • Resource Point definitions
  • Resource Point relationships
  • Accepted Resource Measurements
  • Occupancy snapshots
  • Lease and facility metadata
  • Observation source bindings where needed

Typical tables:

core_metadata.entity_registry
core_relations.entity_relationships
core_metadata.resource_point_registry
core_relations.resource_point_relationships
core_data.accepted_resource_measurements
core_data.occupancy_snapshots
core_metadata.lease_details

7. Outputs​

UOM engines produce:

  • Aggregated resource values
  • Facility-level metrics
  • Building-level metrics
  • Entity-level rollups
  • Resource coverage results
  • Measurement completeness results
  • Reconciliation findings
  • ZARA-ready evidence payloads

All outputs must preserve provenance.

8. Position in the computation chain​

UOM Graph Infrastructure
  → Accepted Measurements
  → UOM Engines
  → ESG Engines
  → Graph-Aware ESG Intelligence
  → ZARA / Reports Hub

UOM engines sit above the physical/resource graph and below higher-order ESG interpretation.

9. Relationship to NORM​

UOM engines must not implement independent unit conversion logic.

When unit conversion is required, they call:

MEID_CORE_NORM

This ensures that all unit normalization is centralized, versioned, and auditable.

10. Relationship to Graph-Aware ESG Intelligence​

UOM engines perform deterministic computations.

Graph-Aware ESG Intelligence orchestrates:

graph traversal
entity discovery
context resolution
MICE selection
result synthesis

The intelligence layer calls UOM engines rather than duplicating their logic.

11. Audit & Provenance​

Every UOM execution should record:

  • engine ID
  • engine version
  • client ID
  • entity IDs used
  • Resource Points used
  • measurement IDs used
  • allocation logic
  • warnings
  • confidence score
  • trust score

UOM engines are integrated with ALTD / DAL logging.

12. Registry view​

All engines tagged as uom:

Loading micro engines…

13. Design rationale​

UOM engines make the organizational graph computational.

They allow ZAYAZ to answer questions such as:

How much electricity did the Denver office use?
Which facilities have no water data?
What is kWh per employee for each office?
Which buildings are resource outliers?

By separating UOM computations into dedicated Micro Engines, ZAYAZ avoids monolithic ESG logic and creates reusable, auditable, graph-aware computation services.

  • NORM — Normalization Engines
  • AGGR — Aggregation Engines
  • VALI — Validator Engines
  • CALC — Calculation Engines
  • SEM — Stochastic Extrapolation Module
  • ESG — ESG attribution and boundary engines
  • ZARA — Explainability and orchestration
Draft


GitHub RepoRequest for Change (RFC)