Technical reference / India / MIT Python library
AION Indian Market Calendar
NSE / BSE / MCX / NSE CDS / NO AION API KEY
Scheduled trading days, segment sessions, settlement dates and selected expiries. Keep the exchange, timezone and data generation beside each answer.
Open-source developer infrastructure, not a hosted predictive model, brokerage API or live exchange-status feed.
One date can have several correct calendar answers.
Trading-day, settlement-day and intraday-phase helpers answer different questions. A date can be open for NSE cash and closed for currency derivatives or settlement. A pre-open session can be active without ordinary continuous matching.
The package is free under MIT, available worldwide for its defined Indian-market scope, and uses bundled or validated cached data without an AION account. One universal wheel serves the tested macOS, Linux and Windows library environments on Python 3.10-3.14; Python 3.9 is unsupported. Online data checking and Python-code installation are separate operations.
NSE currency derivatives use a different clock.
| Segment | Scheduled behavior | Boundary |
|---|---|---|
| NSE / BSE cash | Regular pre-open 09:00-09:15 IST; continuous matching 09:15-15:30, subject to dated rules. | Distinguish generic cash from the F&O-underlying cash subset for closing auctions. |
| NSE CDS INR pairs | USDINR, EURINR, GBPINR and JPYINR: 09:00-17:00 IST, with currency-specific holidays. | No BSE CDS, cross-currency contracts, currency expiry calendar or arbitrary contract cutoff. |
| Equity derivatives | Supported NSE/BSE schedules and selected index expiry helpers. | A generic segment query cannot identify an individual expiring contract cutoff. |
| Non-agri commodities | MCX/NSE seasonal close 23:30 during US daylight saving, otherwise 23:55, subject to sourced overrides. | No agricultural contract-specific hours. Seasonal time calculations are not holiday coverage for another year. |
| Settlement | Separate NSE Clearing-style date utility. | Not a universal banking, BSE or MCX settlement service. |
| Muhurat and expiries | Notified NSE cash/F&O special windows; NIFTY weekly Tuesday, supported NSE monthly indices last Tuesday and SENSEX Thursday, with covered holidays shifted to the preceding trading day. | Unknown NSE F&O special schedules refuse. BSE/MCX Muhurat metadata retains provisional labels. |
On narrow screens, scroll within the table; all rows remain visible vertically.
Example: NSE equity trades on 1 April 2026 while CDS and settlement are closed. Do not reuse the equity answer for a currency or settlement workflow.
Effective 7 December 2026
The sourced rule distinguishes F&O-underlying cash continuous matching to 15:30, the scheduled closing-auction envelope at 15:30-15:40, general equity derivatives to 15:45, and cash post-close at 15:50-16:00. Expiring index contracts stop at 15:30 on their expiry day; the generic segment answer is not an individual contract master.
Phase and continuous-trading windows are half-open. The legacy is_market_open activity envelope retains an inclusive close edge for compatibility; it is not an ordinary-order gate. The auction envelope does not observe random order-entry closure or actual auction completion. No live halt, broker permission, execution or compliance guarantee is provided.
Approved calendar data can change without a new wheel.
Calendar-data flow
- Primary source evidenceExchange, settlement or legal notice; a source observation is not an approved change.
- Publisher reviewReview date, jurisdiction and affected MIC/segment before publishing a versioned dataset.
- Public data channelApproved JSON files with manifest, version and SHA-256 bindings.
- Installed client checkCompatible client makes a bounded online check; offline policy can prevent it.
- Validate and adoptAccept a validated cache generation or retain the verified fallback; report its identity.
The 1.8.0 wheel bundles dataset 2026.11.14. The public Indian in/ manifest currently reports 2026.11.03; do not confuse that older remote generation with the newer bundle, or apply the Lithuanian 2026.11.15.1 identifier to India.
Compatible online checks occur at most daily and validate manifest/schema/payload hashes before atomic adoption. The data manifest is hash-verified, not digitally signed. Offline mode prevents refresh. A rejected refresh keeps validated data and does not invent unsupported dates. Publisher approval and client integrity checks solve different problems: an intact file is not automatically a correct exchange calendar.
A data refresh is not a Python software upgrade.
Independent Python software flow
Normal supported path: the user controls the application interpreter.
- Published package releasePython code and bundled data in an approved PyPI distribution.
- Manual upgrade and verificationUser upgrades the application environment, checks its version and retests. This path ends here: no managed enrollment occurs.
Optional alternative / not production-activated
- Separate explicit enrollmentA user chooses a distinct owned managed runtime. Normal pip never makes this choice.
- Approved trust and policyIndependent signing, channel, bootstrap and compatibility admission are required.
- Isolated patch and rollbackHealth checks and workload controls precede any managed promotion. Production activation remains disabled.
Existing users need an explicit upgrade in the Python environment used by their application. Ordinary pip installation registers no background job, enrolls no managed runtime and does not switch a running application's interpreter.
The 1.8.0 update components passed isolated engineering tests, but production signing/channel/bootstrap approval remains separate. Windows updater rollout is paused. Windows 10 x64, native Windows 11 x64 and real reboot/logon acceptance remain unverified.
Version discovery, notification-only enrollment and installation have different requirements. A 1.7.1 installation cannot receive a new checker retroactively. Release subscribers or the PyPI release feed can follow announcements; PyPI cannot push notices to every earlier installer.
Use the interpreter and timezone your application uses.
python -m pip install --upgrade aion-indian-market-calendar
python -c "import importlib.metadata as m; print(m.version('aion-indian-market-calendar'))"import os
os.environ["AION_IN_CALENDAR_OFFLINE"] = "1"
from aion_indian_market_calendar import (
IndiaMarketCalendar, get_market_phase,
)
cal = IndiaMarketCalendar.bundled(2026)
assert cal.is_market_open(
"2026-11-13T16:30:00+05:30", market="CDS"
)
assert not cal.is_continuous_trading(
"2026-11-13T09:05:00+05:30", market="NSE"
)
assert get_market_phase(
"2026-12-07T15:30:00+05:30",
"NSE_EQUITY_FNO_UNDERLYING", calendar=cal,
) == "CLOSING_AUCTION"Use these bounded answers for exchange-aware schedulers, backtests, event-intelligence pipelines and fintech applications. A date-only trading answer is not a live venue-status check. covers() reports the precise interval; a holiday list for a partial covered year does not establish full-year support. Caller-owned custom records remain the caller's completeness responsibility.
Class methods are datetime-first. The module helper is market-first: is_market_open("CDS", at=...). Aware values convert to IST; naive clocks retain the documented IST compatibility assumption and warn.
The earlier defect remains part of the release record.
Correction for published 1.7.1
Isolated checks reproduced unsupported-date, timezone-date, session-phase, Muhurat and loaded-refresh defects. We apologize for these errors and the uncertainty they can cause. Version 1.8.0 corrects the reproduced cases; a current holiday cache alone does not repair old Python code.
Read the affected-function matrix and manual upgrade guidance. Earlier releases beyond the specifically tested evidence are not assumed affected.
Show us the exact date and the expected answer.
For Indian Market Calendar, email [email protected] or open a source issue. Include:
- Package name and installed version
- Dataset generation from the status helper
- MIC or exchange/segment
- Date, time and explicit timezone
- Expected versus actual answer
- Official source link and a minimal reproducer
Do not send credentials, customer records, account data, private strategies, orders or unnecessary personal information. This is a support contact, not a response-time guarantee.