powernse 0.1.0

Official NSE India archives on disk. One CLI. A small Python API.
Bhavcopy, F&O, indexes, deals, corporate actions — downloaded from the exchange, not scraped from a third-party dump.

Docs · Quick start · PyPI · Changelog

NSE publishes end-of-day files. The URLs move, the session cookie is picky, and nobody wants to re-learn the layout every quarter. PowerNSE stages those archives under a local nse-data/ tree, keeps a download manifest, and lets you query OHLC from the CLI or from Python.

Need a ready-made tree? powernse fetch-bundle pulls the tracked nse-data/ from this GitHub repo. Prefer building it yourself? Point the downloaders at NSE.

Python 3.13+.

Install

pip install powernse
pip install 'powernse[pandas]'   # optional DataFrame helpers

Quick start

powernse bhavcopy --resume
powernse fo-bhavcopy --resume --days 30
powernse index-closes --from 2024-08-01 --to 2024-08-05
powernse fetch-bundle --force    # optional: GitHub-hosted archive
powernse status
powernse ohlc RELIANCE --from 2024-08-01 --to 2024-08-05
powernse doctor

from datetime import date
from powernse import BhavcopyDownloader, NSEData

BhavcopyDownloader("./nse-data").download_range(date(2024, 8, 1), date(2024, 8, 5))
data = NSEData("./nse-data")
bars = data.ohlc("RELIANCE", from_date=date(2024, 8, 1), to_date=date(2024, 8, 5))

Full walkthrough: docs site quickstart.

What you get

Trading days use XBOM sessions when exchange-calendars covers the window (from 2006-08-16). Earlier dates fall back to weekdays.

Archive layout

nse-data/
  raw/bhavcopy/YYYY/YYYY-MM-DD.csv
  raw/fo_bhavcopy/…
  raw/full_bhavcopy/…
  raw/index_closes/…
  raw/bulk_deals/…  raw/block_deals/…  raw/fo_secban/…
  raw/corporate_actions/YYYY/YYYY-MM-DD.json
  raw/index_constituents/YYYY/YYYY-MM-DD_<index>.json
  manifest/downloads.jsonl

Etiquette

Requests are throttled and an NSE session cookie is primed first. Keep date ranges modest; --skip-existing is on by default. Exchange URLs can change without notice.

This project does not download or ingest Kaggle (or other third-party) dumps. If you only need rough historical OHLCV for experiments, fetch those yourself — PowerNSE sticks to official archives.

Disclaimer

PowerNSE is not affiliated with, endorsed by, or connected to the National Stock Exchange of India (NSE) or any related entity. It is an unofficial tool for educational and research use only — not financial advice, and not an official data feed.

You are responsible for how you use the software and any data it retrieves, including compliance with NSE terms of use. The author accepts no liability for losses, damages, or other consequences arising from use of this package. Use at your own risk.

Full text: DISCLAIMER.md.

Documentation

Development

uv sync
make check    # ShipGate
make format
make test
make build

Local docs: uv run --with mkdocs-material mkdocs serve. Quality gates: ShipGate (suite: standard).

License

MIT — see LICENSE. Also read DISCLAIMER.md.