imf-to-neat — NEAT plugin for IMF ontologies

Converts one or more PCA IMF Turtle (.ttl) ontology files into a NEAT physical data model (views, containers, properties with value types).

Features

• Aspect-aware view naming (Plant, Product, Installed, Function, Location)
• Value-type resolution from CFIHOS, RDS (PCA SPARQL), IEC CDD and label heuristics
• Deprecated-term cleaning (dcterms:replaces)
• Container optimization: CFIHOS hierarchy grouping, identical-fingerprint merging, subset detection, AllInstalled pooling

Installation

# From the plugin directory
pip install -e .

# or with uv
uv pip install -e .

The plugin and cognite-neat must be in the same Python environment.

Usage — NEAT plugin API

from cognite.neat import NeatSession

neat = NeatSession()  # or NeatSession(client) with a CDF client

# Single file
neat.plugin.data_model.read(
    "imf",
    io="imf_ontology.ttl",
    clean=True,
    optimize_containers=True,
    space="my_imf_space",
    external_id="MyIMFDataModel",
    version="v1",
)

# Multiple files merged into one model
neat.plugin.data_model.read(
    "imf",
    io=["types_a.ttl", "types_b.ttl"],
    cfihos_csv="CFIHOS CORE property v2.0.csv",
)

All keyword arguments

Usage — standalone CLI

After installation the imf-to-neat command is available:

imf-to-neat imf_ontology.ttl --clean --optimize-containers
imf-to-neat a.ttl b.ttl -o merged_model.yaml
imf-to-neat ontology.ttl --cfihos-csv "CFIHOS CORE property v2.0.csv" --rds-sparql

Plugin structure

imf-to-neat-plugin/
├── pyproject.toml
├── README.md
└── imf_to_neat/
    ├── __init__.py
    ├── plugin.py       # IMFToNeatPlugin (DataModelImporterPlugin)
    ├── _importer.py    # IMFToNeatImporter (BaseImporter)
    ├── _core.py        # All processing logic
    └── cli.py          # CLI entry point (imf-to-neat command)