gridmodel-network-analysis

Station-to-station EHV analysis exports plus full raw graph dumps derived from the local gridmodel-datasets seed outputs.

Layout

• wave1/
  • curated 500 kV + unified 220/230 kV station graph
  • plain JSON for direct inspection in GitHub
• full_raw/
  • compressed JSONL dumps for raw nodes, raw edges, seed components, attrs, connections, and membership
  • full_raw/*.gz tracked with Git LFS
• naming/
  • substation and line context packs for human/LLM-assisted naming research
  • naming/*.gz tracked with Git LFS
  • proposed names must come back as candidates with source URLs and confidence, not direct truth edits
• scripts/build_wave1_exports.py
  • generator and validator
• scripts/build_naming_pack.py
  • generator for substation/line naming research artifacts

Source Data

By default the generator reads from the sibling repository:

• /Users/nikolaybubnov/Downloads/gridmodel-dev/gridmodel-datasets/08_Data/08.01_Seed/scripts/output

You can override that with either:

• GRIDMODEL_DATASET_OUTPUT_ROOT=/abs/path/to/output
• GRIDMODEL_DATASETS_ROOT=/abs/path/to/gridmodel-datasets

Generate

python3 scripts/build_wave1_exports.py
python3 scripts/build_naming_pack.py

The script:

1. builds the curated Wave 1 station graph
2. emits the full raw dumps
3. writes a deterministic manifest
4. validates the exported contracts before exiting

See wave1/README.md for the curated graph contract.

See naming/README.md for the naming research workflow and candidate output contract.