ODS-E & Architecture

ODS-E (Open Data Schema for Energy) is an open specification for interoperable energy asset data across generation, consumption, and net metering. It ships with versioned JSON schemas, vendor transforms, a Python reference runtime with CLI, and integration guides for inverters, industrial protocols, SCADA, and utility data portals.

The Problem

Every OEM — Huawei, Enphase, SolarEdge, Fronius, SMA, Vestas, and dozens more — exposes telemetry through proprietary APIs, CSV exports, and SCADA protocols with different field names, units, and error taxonomies. Integrating each new vendor typically costs 120–160 hours of engineering effort: mapping columns, normalizing timestamps, classifying error states, and writing bespoke validation. With 20+ common OEMs across solar PV, wind, BESS, and utility metering, the integration tax compounds rapidly.

ODS-E eliminates this by defining a single canonical schema. OEM-specific data is transformed into ODS-E records via declarative YAML specs, validated against versioned schemas, and then consumed by any downstream system — analytics platforms, ML pipelines, billing engines, or market settlement workflows — without re-integration.

Components

JSON Schemas

Versioned JSON Schema (Draft 2020-12) definitions for energy data structures:

• energy-timeseries.json — 65-field canonical record for a single energy reading (3 required fields, 62 optional). Covers core telemetry, market settlement, wheeling, tariffs, curtailment, BESS dispatch, wind SCADA, and renewable certificates.
• asset-metadata.json — asset configuration, location, and grid connection details.
• equipment-register.json — equipment hierarchy and technical attributes from ERP/EAM systems.
• 5 additional schemas for equipment ID mapping, maintenance history, spare parts, failure taxonomy, and more.

Error Taxonomy

Seven standardized error categories that replace OEM-specific error codes:

Transform Specifications

Declarative YAML files that map OEM-specific data to ODS-E records. Each spec defines input column aliases, output field mappings, error code mappings, and validation bounds. 20 vendor transforms are included — see Transform Specifications.

Python Reference Runtime + CLI

A Python package (odse) providing:

• transform() — convert vendor data to ODS-E records
• validate() / validate_batch() — validate records against schemas and conformance profiles
• to_json() / to_parquet() — persist validated records
• CLI commands: odse transform and odse validate

Quick Start

Install

pip install -e src/python[parquet]

Transform vendor data

odse transform --source generic_csv \
  --input examples/fixtures/quickstart_scada.csv \
  --column-map timestamp=Timestamp,kWh=ActiveEnergy,asset_id=Asset \
  --output examples/output/quickstart.cleaned.json

Validate

odse validate --input examples/output/quickstart.cleaned.json --level schema

Expected output:

3/3 valid, 0 errors

In Python

from pathlib import Path
from odse import transform, validate_batch, to_parquet

records = transform(
    Path("examples/fixtures/quickstart_scada.csv"),
    source="generic_csv",
    column_map={
        "timestamp": "Timestamp",
        "kWh": "ActiveEnergy",
        "asset_id": "Asset",
    },
)

result = validate_batch(records, level="schema")
print(result.summary)
# 3/3 valid, 0 errors

to_parquet(records, "output/", partition_by=["asset_id", "year", "month", "day"])

What’s Next

• Ona Intelligence Layer Architecture — how ODS-E fits within the broader platform
• ODS-E & the SDK — using ODS-E with the Ona SDK
• Schema Reference — field-level documentation for all schemas
• Transform Specifications — vendor transform catalog and usage
• Validation Guide — validation levels and conformance profiles

Repository

Source code and schemas: github.com/AsobaCloud/odse

Licensing:

• Specification, schemas, transforms: CC-BY-SA 4.0
• Reference implementation and tools: Apache 2.0

Maintained by Asoba Corporation.