aynse 2.3.0

A lean and modern Python library to fetch data from NSE (National Stock Exchange of India)

aynse

aynse is a lean, modern python library for fetching data from the national stock exchange (nse) of india. it includes a resilient http client (http/2, retries with jitter, connection pooling), adaptive batching, and efficient streaming utilities.

features

• historical data: canonical stock, index, derivative, and index valuation records
• archive datasets: bhavcopy, full bhavcopy, F&O bhavcopy, index bhavcopy, bulk deals, and index constituents
• live market data: standardized quotes, option chains
• cli: simple commands for quick downloads
• resilient networking: http/2, connection pooling, retries with exponential backoff, rate limiting, circuit breaker
• batching & streaming: adaptive concurrency and low-memory processing
• comprehensive type hints: full typing support for ide autocomplete
• extensive test coverage: robust test suite for reliability
• streaming utilities: chunked CSV/JSON/ZIP processing for larger files

installation

install aynse directly from pypi:

pip install aynse

for development:

pip install aynse[dev]
# or
pip install -r requirements.dev.txt

etymology / musings

aynse is a portmanteau of "ayn" from miss ayn rand and "nse" from national stock exchange. ayn rand was a russian-american writer and philosopher known for her philosophy of objectivism, which emphasizes individualism and rational self-interest. among other things, she was a strong advocate for laissez-faire capitalism.

the name serves as a fun ironical reminder of the library's purpose: to provide a tool for individuals to access and analyze financial data independently, without relying on large institutions or complex systems.

in a cruel twist of fate, this open source library wouldn't be encouraged under ayn rand's philosophy, as she discouraged altruism and believed in the pursuit of one's own happiness as the highest moral purpose.

and as the final act of irony, we gather to use this library to analyze financial markets while generating zero (and possibly, negative) intrinsic value for humankind as a whole - this is capitalism.

quick start

get historical stock data

retrieve historical data for a stock as a pandas dataframe:

from aynse import stock_df

df = stock_df(
    symbol="reliance",
    from_date="2024-01-01",
    to_date="2024-01-31",
)

print(df[["date", "symbol", "open", "close"]].head())

archive datasets

from aynse import bhavcopy_raw, bhavcopy_save

records = bhavcopy_raw("2024-07-26")
print(records[0])

path = bhavcopy_save("2024-07-26", "downloads/")
print(path)

get live stock quote

fetch live price information for a stock:

from aynse import NSELive

live = NSELive()
quote = live.stock_quote("INFY")

print(quote["symbol"])
print(quote["company_name"])
print(quote["price"]["last"])
print(quote["price"]["change_percent"])

get option chain data

fetch option chain for index or equity:

from aynse import NSELive, summarize_option_chain

live = NSELive()
chain = live.equities_option_chain("RELIANCE")
summary = summarize_option_chain(chain)

print(summary["at_the_money_strike"])
print(summary["put_call_ratio"])
print(summary["max_pain"])
print(summary["at_the_money_implied_volatility"])

metadata

from aynse import dataset_capabilities, supported_indices

print(dataset_capabilities()["historical"]["outputs"])
print(supported_indices()[:3])

canonical contracts

accepted inputs

All public date boundaries accept:

• datetime.date
• datetime.datetime
• "YYYY-MM-DD"

Symbols are normalized at the boundary, so "reliance" and "RELIANCE" resolve the same way for symbol-based APIs.

historical record shape

stock_raw(...) returns rows like:

{
    "date": date(2024, 1, 1),
    "symbol": "RELIANCE",
    "series": "EQ",
    "open": 2850.0,
    "high": 2875.0,
    "low": 2832.0,
    "close": 2868.0,
    "previous_close": 2844.0,
    "last_traded_price": 2867.5,
    "volume": 1234567,
}

Time-series datasets are returned in chronological ascending order.

archive contract

Archive families follow the same triplet:

• bhavcopy_raw(...), bhavcopy_df(...), bhavcopy_save(...)
• full_bhavcopy_raw(...), full_bhavcopy_df(...), full_bhavcopy_save(...)
• bhavcopy_fo_raw(...), bhavcopy_fo_df(...), bhavcopy_fo_save(...)
• bhavcopy_index_raw(...), bhavcopy_index_df(...), bhavcopy_index_save(...)
• bulk_deals_raw(...), bulk_deals_df(...), bulk_deals_save(...)
• index_constituent_raw(...), index_constituent_df(...), index_constituent_save(...)

live contract

live methods return stable top-level payloads. For example:

{
    "symbol": "INFY",
    "company_name": "Infosys Limited",
    "price": {
        "last": 1520.5,
        "change": 12.4,
        "change_percent": 0.82,
    },
}

Option-chain methods return:

• symbol
• market_type
• timestamp
• underlying_value
• expiry_dates
• strike_prices
• records
• summary

The summary includes call/put open interest, changes in open interest, traded volume, PCR, ATM strike and implied volatility, strongest call/put walls, and the open-interest-weighted max-pain strike.

public API highlights

historical + archives

• stock_raw, stock_df, stock_csv
• index_raw, index_df, index_csv
• index_pe_raw, index_pe_df
• derivatives_raw, derivatives_df, derivatives_csv
• bhavcopy_*, full_bhavcopy_*, bhavcopy_fo_*, bhavcopy_index_*
• bulk_deals_*
• index_constituent_*
• expiry_dates

live + metadata

• NSELive.stock_quote, stock_quote_fno, trade_info, market_status
• NSELive.option_chain_contract_info, index_option_chain, equities_option_chain, currency_option_chain
• NSELive.corporate_announcements, corporate_actions, results_calendar
• NSELive.metadata
• dataset_capabilities, supported_indices, supported_instruments, supported_event_categories

analytics

• add_returns
• add_rolling_volatility (configurable annualization period)
• add_moving_average (simple or exponential)
• add_rsi
• add_atr
• add_bollinger_bands
• add_drawdown
• add_gap_metrics
• add_volume_metrics
• summarize_option_chain
• analyze_event_window

CLI

# historical stock data
aynse stock -s RELIANCE -f 2024-01-01 -t 2024-03-31 -o reliance_q1.csv

# live quote
aynse quote -s RELIANCE

# archive download
aynse bhavcopy -d downloads/ -f 2024-07-26

# holidays
aynse holidays -y 2024

testing

# deterministic/offline contracts
pytest -m offline -v --tb=short

# live NSE/RBI integration checks
pytest -m live -v --tb=short

# end-to-end integration checks
pytest -m e2e -v --tb=short

# full suite (offline + live)
pytest tests -v --tb=short

release workflow

releases are automated after the test workflow succeeds on master.

1. make your changes and bump the version

make bump-patch  # or bump-minor, bump-major

2. commit and push to master

git commit -am "Bump version to X.Y.Z"
git push

if X.Y.Z is not already published, github actions builds and verifies the wheel and source distribution, publishes both to pypi, creates tag vX.Y.Z, and creates the github release. live NSE tests remain advisory because upstream availability is outside the package's control.

everything to do with release/version control

show current version

python scripts/bump_version.py --current

bump versions (updates pyproject.toml only)

python scripts/bump_version.py patch # 1.1.0 -> 1.1.1 python scripts/bump_version.py minor # 1.1.0 -> 1.2.0 python scripts/bump_version.py major # 1.1.0 -> 2.0.0 python scripts/bump_version.py --set 2.0.0 # set exact version

preview changes without modifying

python scripts/bump_version.py patch --dry-run

documentation

• Historical Data Guide
• Live Data Guide
• API Reference
• Data Calls Reference

contributing

contributions are welcome. Please add or update contract tests whenever you change a public schema, accepted input, or save behavior.

license

this project has a (custom) mit* license but extends limitations. if you're an agency/corporate with >2 employees, you cannot wrap this project or use it without prior written permission from the author. if you're an individual, you can use it freely for personal projects.

this project is not intended for commercial use without prior permission. if caught using this project in violation of the license, it may result in automated reporting and/or legal action.

please see the license file for more details.