Powergrid-India

Author: DK Mendiratta

Unofficial public skill and workbook builder for PAN India renewable-energy evacuation substations on the CTU / PGCIL grid. This is not POWERGRID, PGCIL, or CTUIL software.

Canonical source: https://github.com/d33pm3/Powergrid-India

This is / this is not

This is an unofficial planning workbook builder and AI skill for PAN-India RE evacuation substations. This is a 132-row public baseline plus an optional live reconcile against public CTUIL / PGCIL / CEA / MNRE / GEC pages. This is a local Python package that writes a 10-sheet .xlsx under outputs/. This is not POWERGRID, PGCIL, or CTUIL official software. This is not a verified connectivity, GNA/LTA, or CEA regulatory filing. This is not a web app, API, Docker stack, or always-on scraper. This is not tariff, discom, or thermal-plant modelling. This is not a live-refreshed dataset in a fresh clone — the CLI uses the embedded baseline only.

The 132-record baseline is planning-grade, compiled from public reporting. Capacity, MVA, and developer-allocation figures are indicative. Rows that are not confirmed against a live source in the current session must be tagged [SOURCE_FALLBACK:BASELINE] and must not be described as verified regulatory filings.

Who it is for

• RE planners mapping evacuation nodes and corridor capacity
• InvIT / transmission analysts who need a planning-grade master, not a filing
• GRC advisors checking source tags and planning vs verified claims
• Agent builders who will load SKILL.md and run the live-research path

What it does

• Loads a 132-substation master (5 regions, 30 states, ~130,500 MW baseline)
• Optionally merges live updates from public CTUIL / PGCIL / CEA / MNRE / GEC sources (agent path only)
• Builds a 10-sheet .xlsx workbook (9 data sheets + an 8-block Analytics Dashboard)

Workbook: 10 sheets and 11 columns

Sheets: All_Substations, Regional_Summary, State_Summary, Western_Region, Northern_Region, Southern_Region, Eastern_Region, NorthEastern_Region, Top_50_Substations, Analytics_Dashboard.

Columns on the master: Region, State, Substation_Name, Type_AIS_GIS, Voltage_Level, Transformation_Capacity_MVA, Renewable_Companies_Connected, Project_Type, Total_Capacity_MW, Remarks, Data_Source.

Repository layout

Powergrid-India/
├── SKILL.md                         Skill workflow, schema, QA gates
├── references/
│   ├── substation-master.md         132-record baseline notes
│   ├── substations-a.csv            Baseline rows 1–66
│   ├── substations-b.csv            Baseline rows 67–132
│   └── generation-script.md         openpyxl builder reference
├── evals/                           Trigger and output eval cases
├── src/powergrid_india/             Runnable package
├── tests/                           Parser, reconcile, workbook tests
├── outputs/                         Local artefacts (gitignored)
├── LICENSE
├── SECURITY.md
└── pyproject.toml

This repository is a new public build. It is not linked to any private catalogue or private extractor repo.

Use the baseline CLI (no live fetch)

This path does not call CTUIL, PGCIL, CEA, MNRE, or GEC. It builds the workbook from the embedded substations-a.csv + substations-b.csv baseline only. It is not a server and is not deployed as a web app.

Requires Python 3.10+ and openpyxl.

git clone https://github.com/d33pm3/Powergrid-India.git
cd Powergrid-India
python3 -m venv .venv && source .venv/bin/activate
python -m pip install -e ".[dev]"
mkdir -p outputs
powergrid-india --output outputs/PAN_India_RE_Substations_Master_Data.xlsx

Equivalent module form:

python -m powergrid_india.cli --output outputs/PAN_India_RE_Substations_Master_Data.xlsx

Run live research in an agent

Use this path when you need a refreshed workbook, not the frozen baseline.

1. Clone the repo (same commands as above, including pip install -e ".[dev]").
2. Drop SKILL.md into the skill host (Claude Skills, Cursor, Grok custom skills, or equivalent).
3. Give the agent live web search and page-fetch. Point it at the five public URLs below.
4. The agent reconciles live hits against the 132-row baseline, tags [SOURCE_LIVE] or [SOURCE_FALLBACK:BASELINE], then writes the same 10-sheet workbook under outputs/.

The CLI does not perform that live merge. Do not describe a baseline-only xlsx as a live-verified inventory.

Tests

pytest -q

Public sources (live-research targets)

• https://www.ctuil.in
• https://www.powergrid.in
• https://www.cea.nic.in
• https://www.mnre.gov.in
• https://www.powergrid.in/project/green-energy-corridor

License

MIT. See LICENSE.

You may use this code; the xlsx is not a filing.