re https://github.com/rsheftel/pandas_market_calendars/issues/41

Summary status

• [x] XHKG
  • https://github.com/quantopian/trading_calendars/issues/152#issuecomment-689799967
• [x] CFE
  • https://github.com/quantopian/trading_calendars/issues/152#issuecomment-689812670
  • Note: We don't know what to do about 1994-04-27 in pmc but not in tc
• [ ] XSWX
• [X] LSE
  • Issue: https://github.com/quantopian/trading_calendars/issues/163
  • PR: https://github.com/quantopian/trading_calendars/pull/166
• [ ] XSHG
• [ ] XASX
• [X] TSX
  • https://github.com/quantopian/trading_calendars/issues/152#issuecomment-689136700)
• [ ] BMF

Extras not covered by diff below:

• [x] Pre 1990 NYSE holidays
  • https://github.com/quantopian/trading_calendars/issues/162

Details

code generating the table below

from collections import namedtuple 

import pandas_market_calendars as pmc
import trading_calendars as tc

# todo port this back to trading calendars
def get_tc_calendar_names():
  return list(tc.calendar_utils._default_calendar_factories.keys())+list(tc.calendar_utils._default_calendar_aliases.keys())

tc.get_calendar_names = get_tc_calendar_names
# end todo

pmc_calendars = set(pmc.get_calendar_names())
tc_calendars = set(tc.get_calendar_names())
tc_to_pmc_mapping = {
    'XASX': 'ASX',
    'XHKG': 'HKEX',
    'XTKS': 'JPX',
    'XOSL': 'OSE',
    'XSWX': 'SIX',
    'XSHG': 'SSE',
}

in_both = (tc_calendars & pmc_calendars) | set(tc_to_pmc_mapping.keys())

CompareResult = namedtuple("Result", ["only_tc", "only_pmc"])

def compare_calendars(cal: str):
  tc_cal = tc.get_calendar(cal).all_sessions
  pmc_cal_name = tc_to_pmc_mapping.get(cal, cal)
  pmc_cal = pmc.get_calendar(pmc_cal_name).valid_days(min(tc_cal), max(tc_cal))
  if not (tc_cal ^ pmc_cal).empty:
    return CompareResult

out = {}
for cal in in_both:
  diff = compare_calendars(cal)
  if diff:
    out[cal] = diff

# everything below this is markdown formatting
out_str = []
for cal, result in out.items():
  out_str.append(f"## {cal} \n")
  out_str.append("Trading Days only in `trading_calendars` \n")
  out_str.extend([f"- {item} \n" for item in result.only_tc])
  out_str.append("\n")
  out_str.append("Trading Days only in `pandas_market_calendars` \n")
  out_str.extend([f"- {item} \n" for item in result.only_pmc])
  out_str.append("\n")
md = "".join(out_str); print(md)

XHKG

Trading Days only in trading_calendars

• 1990-06-18 00:00:00+00:00
• 1990-08-27 00:00:00+00:00
• 1991-02-14 00:00:00+00:00
• 1991-06-18 00:00:00+00:00
• 1991-08-26 00:00:00+00:00
• 1992-06-15 00:00:00+00:00
• 1992-07-22 00:00:00+00:00
• 1992-08-31 00:00:00+00:00
• 1993-06-14 00:00:00+00:00
• 1993-08-30 00:00:00+00:00
• 1993-09-17 00:00:00+00:00
• 1994-06-14 00:00:00+00:00
• 1994-08-29 00:00:00+00:00
• 1995-06-19 00:00:00+00:00
• 1995-08-28 00:00:00+00:00
• 1996-06-17 00:00:00+00:00
• 1996-08-26 00:00:00+00:00
• 1997-02-06 00:00:00+00:00
• 1997-06-30 00:00:00+00:00
• 1997-07-02 00:00:00+00:00
• 1997-08-18 00:00:00+00:00
• 1997-10-02 00:00:00+00:00
• 1998-08-17 00:00:00+00:00
• 1998-10-02 00:00:00+00:00
• 1999-09-16 00:00:00+00:00
• 1999-12-31 00:00:00+00:00
• 2001-07-06 00:00:00+00:00
• 2001-07-25 00:00:00+00:00

Trading Days only in pandas_market_calendars

• 1990-05-01 00:00:00+00:00
• 1990-05-02 00:00:00+00:00
• 1990-07-02 00:00:00+00:00
• 1990-10-01 00:00:00+00:00
• 1991-05-01 00:00:00+00:00
• 1991-05-21 00:00:00+00:00
• 1991-07-01 00:00:00+00:00
• 1991-10-01 00:00:00+00:00
• 1992-05-01 00:00:00+00:00
• 1992-05-11 00:00:00+00:00
• 1992-07-01 00:00:00+00:00
• 1992-10-01 00:00:00+00:00
• 1993-05-28 00:00:00+00:00
• 1993-07-01 00:00:00+00:00
• 1994-05-02 00:00:00+00:00
• 1994-05-18 00:00:00+00:00
• 1994-07-01 00:00:00+00:00
• 1995-05-01 00:00:00+00:00
• 1995-05-08 00:00:00+00:00
• 1995-10-02 00:00:00+00:00
• 1996-05-01 00:00:00+00:00
• 1996-05-24 00:00:00+00:00
• 1996-07-01 00:00:00+00:00
• 1996-10-01 00:00:00+00:00
• 1997-05-01 00:00:00+00:00
• 1997-05-14 00:00:00+00:00
• 1998-05-01 00:00:00+00:00
• 1998-05-04 00:00:00+00:00
• 2021-04-06 00:00:00+00:00

CFE

Trading Days only in trading_calendars

Trading Days only in pandas_market_calendars

• 1994-04-27 00:00:00+00:00
• 2004-06-11 00:00:00+00:00
• 2007-01-02 00:00:00+00:00
• 2012-10-29 00:00:00+00:00
• 2012-10-30 00:00:00+00:00
• 2018-12-05 00:00:00+00:00

XSWX

Trading Days only in trading_calendars

• 1994-12-30 00:00:00+00:00
• 1995-12-29 00:00:00+00:00
• 2000-12-29 00:00:00+00:00
• 2005-12-30 00:00:00+00:00
• 2006-12-29 00:00:00+00:00
• 2011-12-30 00:00:00+00:00
• 2016-12-30 00:00:00+00:00
• 2017-12-29 00:00:00+00:00

Trading Days only in pandas_market_calendars

LSE

Trading Days only in trading_calendars

• 1995-05-08 00:00:00+00:00
• 1999-12-31 00:00:00+00:00

Trading Days only in pandas_market_calendars

• 1995-05-01 00:00:00+00:00

XSHG

Trading Days only in trading_calendars

• 2020-01-31 00:00:00+00:00
• 2022-01-31 00:00:00+00:00
• 2022-04-04 00:00:00+00:00
• 2022-05-02 00:00:00+00:00
• 2022-09-12 00:00:00+00:00
• 2022-10-06 00:00:00+00:00
• 2022-10-07 00:00:00+00:00
• 2023-06-23 00:00:00+00:00
• 2023-10-06 00:00:00+00:00
• 2024-02-09 00:00:00+00:00
• 2024-04-05 00:00:00+00:00
• 2024-09-16 00:00:00+00:00
• 2024-10-07 00:00:00+00:00
• 2025-01-28 00:00:00+00:00
• 2025-05-02 00:00:00+00:00
• 2025-06-02 00:00:00+00:00
• 2025-10-06 00:00:00+00:00
• 2025-10-07 00:00:00+00:00

Trading Days only in pandas_market_calendars

• 2020-10-08 00:00:00+00:00

XASX

Trading Days only in trading_calendars

• 1993-01-27 00:00:00+00:00
• 1994-01-27 00:00:00+00:00
• 1995-01-27 00:00:00+00:00
• 1996-01-29 00:00:00+00:00
• 1998-01-27 00:00:00+00:00
• 1999-01-27 00:00:00+00:00
• 2000-01-27 00:00:00+00:00
• 2001-01-29 00:00:00+00:00
• 2004-01-27 00:00:00+00:00
• 2005-01-27 00:00:00+00:00
• 2006-01-27 00:00:00+00:00
• 2007-01-29 00:00:00+00:00
• 2009-01-27 00:00:00+00:00
• 2010-01-27 00:00:00+00:00
• 2011-01-27 00:00:00+00:00
• 2012-01-27 00:00:00+00:00
• 2015-01-27 00:00:00+00:00
• 2016-01-27 00:00:00+00:00
• 2017-01-27 00:00:00+00:00
• 2018-01-29 00:00:00+00:00
• 2021-01-27 00:00:00+00:00

Trading Days only in pandas_market_calendars

• 1990-12-31 00:00:00+00:00
• 1994-01-26 00:00:00+00:00
• 1994-12-27 00:00:00+00:00
• 1995-01-26 00:00:00+00:00
• 1996-01-26 00:00:00+00:00
• 1998-01-26 00:00:00+00:00
• 1999-01-26 00:00:00+00:00
• 1999-12-31 00:00:00+00:00
• 2000-01-26 00:00:00+00:00
• 2001-01-26 00:00:00+00:00
• 2004-01-26 00:00:00+00:00
• 2005-01-26 00:00:00+00:00
• 2005-12-27 00:00:00+00:00
• 2006-01-26 00:00:00+00:00
• 2007-01-26 00:00:00+00:00
• 2009-01-26 00:00:00+00:00
• 2010-01-26 00:00:00+00:00
• 2010-04-26 00:00:00+00:00
• 2011-01-26 00:00:00+00:00
• 2011-04-26 00:00:00+00:00
• 2011-12-27 00:00:00+00:00
• 2012-01-26 00:00:00+00:00
• 2015-01-26 00:00:00+00:00
• 2016-01-26 00:00:00+00:00
• 2016-12-27 00:00:00+00:00
• 2017-01-26 00:00:00+00:00
• 2018-01-26 00:00:00+00:00
• 2021-01-26 00:00:00+00:00

TSX

Trading Days only in trading_calendars

• 1993-12-24 00:00:00+00:00
• 1999-12-24 00:00:00+00:00
• 2004-12-24 00:00:00+00:00
• 2010-12-24 00:00:00+00:00

Trading Days only in pandas_market_calendars

• 2001-09-11 00:00:00+00:00
• 2001-09-12 00:00:00+00:00

BMF

Trading Days only in trading_calendars

• 1997-07-09 00:00:00+00:00
• 2020-07-09 00:00:00+00:00

Trading Days only in pandas_market_calendars

• 1995-12-29 00:00:00+00:00
• 2000-12-29 00:00:00+00:00
• 2006-12-29 00:00:00+00:00
• 2014-06-12 00:00:00+00:00
• 2017-12-29 00:00:00+00:00

This PR is to add a calendar for Tel Aviv Stock Exchange (TASE) under the operating MIC XTAE.

This is based on the respective calendar from pandas-market-calendars, but adding in holidays for 2022 and 2023 as well. I've reconciled the dates for 2019 to 2023 against the official list; see https://info.tase.co.il/eng/about_tase/corporate/pages/vacation_schedule.aspx.

Open points:

• Unclear which times to use as opening and close times as they differ per market segment; see https://info.tase.co.il/eng/trading/trading_schedule/pages/trading_schedule.aspx. Please advise.
• Different close time on Sundays not yet reflected.
• Since I don't see how holiday rules could be handled with dates in the Hebrew calendar, all holidays are defined as ad-hoc holidays. Is there a way to assign them a name, too, like for regular holidays?
• Unit tests currently only cover 2019, more should be added.

Updates XLON to include holidays from pandas_market_calendars.

Closes https://github.com/quantopian/trading_calendars/issues/163.

See details from original PR: https://github.com/rsheftel/pandas_market_calendars/issues/84

This PR adds proper rules for regular XTAE holidays as opposed to just having everything defined as ad-hoc holidays.

I've done my best to cross-check this against the official information here: https://info.tase.co.il/eng/about_tase/corporate/pages/vacation_schedule.aspx

One open point is that it seems that Shushan Purim (which is one day after Purim and is typically not observed in Tel Aviv, but e.g. Jerusalem) will be observed from 2022 on, but wasn't observed before. I'm waiting for clarification back from TASE on this point.

Nevertheless, the current version of the code is consistent with what is currently listed on the website at least for 2020 - 2023. I also briefly manually checked a few years before, but no unit tests are there for these yet.

On a technical note, I've re-used Pandas Easter class -- which is a subclass of DateOffset -- for all holidays since they likewise fall on different Gregorian dates every year. To be able to do that, I had to import a few (internal) Cython functions and classes. A pure Python or even Cython re-implementation seemed to risky in terms of likely forgetting to handle this or that corner-case. We just have to be aware of these imports since they may introduce some more coupling on the specific Pandas version used.

Let me know your thoughts...

Hi, I think I may have found an error with the start and end times for XNYS when using .is_open_on_minute to return a boolean.

Reading through the core trading session hours for the NYSE, it seems to me that on a regular trading day the market starts at exactly 9:30:00 EST (inclusive) and is closed at exactly 16:00:00 EST (inclusive). I wrote some code to test some times and I am getting conflicting answers.

Currently,

• 9:30:00 returns market closed. I think this should be open.
• 9:30:01 returns market closed. I think this should be open.
• 9:31:00 is the first time returned as open.
• 16:00:00 returns market open. I think this should be closed.

My code attached.

Thanks in advance.

import pandas as pd
import pytz
import trading_calendars

NY = 'America/New_York'
us_calendar = trading_calendars.get_calendar('XNYS')

#Market on Friday at 9:30:00 EST
pd_dt1 = pd.Timestamp(year=2020, month=10, day=9, hour=9, minute=30, second=0, tz=pytz.timezone(NY))
open_boolean1 = us_calendar.is_open_on_minute(pd_dt1)
print("Friday at {} EST. open_boolean should be True: {}".format(pd_dt1.strftime('%H:%M:%S'), open_boolean1))

#Market on Friday at 9:30:01 EST
pd_dt2 = pd.Timestamp(year=2020, month=10, day=9, hour=9, minute=30, second=1, tz=pytz.timezone(NY))
open_boolean2 = us_calendar.is_open_on_minute(pd_dt2)
print("Friday at {} EST. open_boolean should be True: {}".format(pd_dt2.strftime('%H:%M:%S'), open_boolean2))

#Market on Friday at 9:31:0 EST
pd_dt3 = pd.Timestamp(year=2020, month=10, day=9, hour=9, minute=31, second=0, tz=pytz.timezone(NY))
open_boolean3 = us_calendar.is_open_on_minute(pd_dt3)
print("Friday at {} EST. open_boolean should be True: {}".format(pd_dt3.strftime('%H:%M:%S'), open_boolean3))

#Market on Friday at 16:00:00 EST
pd_dt4 = pd.Timestamp(year=2020, month=10, day=9, hour=16, minute=0, second=0, tz=pytz.timezone(NY))
open_boolean4 = us_calendar.is_open_on_minute(pd_dt4)
print("Friday at {} EST. open_boolean should be False: {}".format(pd_dt4.strftime('%H:%M:%S'), open_boolean4))

#Market on Friday at 16:00:01 EST
pd_dt5 = pd.Timestamp(year=2020, month=10, day=9, hour=16, minute=0, second=1, tz=pytz.timezone(NY))
open_boolean5 = us_calendar.is_open_on_minute(pd_dt5)
print("Friday at {} EST. open_boolean should be False: {}".format(pd_dt5.strftime('%H:%M:%S'), open_boolean5))

#Market on Saturday at 9:30:00 EST - Should be closed
pd_dt6 = pd.Timestamp(year=2020, month=10, day=10, hour=9, minute=30, second=0, tz=pytz.timezone(NY))
open_boolean6 = us_calendar.is_open_on_minute(pd_dt6)
print("Saturday at {} EST. open_boolean should be False: {}".format(pd_dt6.strftime('%H:%M:%S'), open_boolean6))

Hello, I'm trying to exclude a business day, any week, from a trading calendar.

Lets say I don't want to trade on any Friday, what I'm doing is changing the weekmask:

class CustomCalendar(IEPAExchangeCalendar):
    def __init__(self):
        super().__init__()

    @property
    def weekmask(self):
       return '111000'

custom_calendar = CustomCalendar()

but I'm getting:

ValueError: Special dates [Timestamp('1992-07-03 17:00:00+0000', tz='UTC'), ..., ...] are not trading days.

What is the best way to exclude business day every week?

I noticed that the load time for most all calendars are around 1.5 seconds, except the XTKS calendar which takes 12.2 seconds to load. Is there any particular reason for that?

This is the code and results I used:

from datetime import datetime
import trading_calendars as tc

def load_times():
    for name in tc.calendar_utils._default_calendar_factories:
        print(name, end='')
        start = datetime.now()
        _ = mcal.get_calendar(name)
        end = datetime.now()
        print(f': run time is {end - start}')

ASEX: run time is 0:00:01.462395
BVMF: run time is 0:00:01.608703
CMES: run time is 0:00:01.247221
IEPA: run time is 0:00:01.221055
XAMS: run time is 0:00:01.239995
XASX: run time is 0:00:01.405535
XBKK: run time is 0:00:01.483120
XBOG: run time is 0:00:01.902455
XBOM: run time is 0:00:00.635975
XBRU: run time is 0:00:01.259231
XBUD: run time is 0:00:01.545929
XBUE: run time is 0:00:01.647241
XCBF: run time is 0:00:01.397777
XCSE: run time is 0:00:01.523843
XDUB: run time is 0:00:01.301222
XFRA: run time is 0:00:01.396143
XHEL: run time is 0:00:01.423694
XHKG: run time is 0:00:01.365153
XICE: run time is 0:00:01.561660
XIDX: run time is 0:00:01.219812
XIST: run time is 0:00:01.244760
XJSE: run time is 0:00:01.435470
XKAR: run time is 0:00:01.075075
XKLS: run time is 0:00:01.191726
XKRX: run time is 0:00:00.940299
XLIM: run time is 0:00:01.465012
XLIS: run time is 0:00:01.357335
XLON: run time is 0:00:01.407390
XMAD: run time is 0:00:01.282585
XMEX: run time is 0:00:01.377445
XMIL: run time is 0:00:01.336494
XMOS: run time is 0:00:01.307830
XNYS: run time is 0:00:01.408174
XNZE: run time is 0:00:01.423343
XOSL: run time is 0:00:01.464810
XPAR: run time is 0:00:01.283910
XPHS: run time is 0:00:01.333504
XPRA: run time is 0:00:01.456183
XSES: run time is 0:00:00.961805
XSGO: run time is 0:00:01.739328
XSHG: run time is 0:00:00.709314
XSTO: run time is 0:00:01.572027
XSWX: run time is 0:00:01.435308
XTAI: run time is 0:00:01.137340
XTKS: run time is 0:00:12.211477
XTSE: run time is 0:00:01.481504
XWAR: run time is 0:00:01.544426
XWBO: run time is 0:00:01.442748
us_futures: run time is 0:00:00.859409
24/7: run time is 0:00:01.463259
24/5: run time is 0:00:01.034509

Adds break_start_times and break_end_times to TradingCalendar to model intraday lunch breaks. These work as vectors of values, similarly to how open_times and close_times work so we can model break times changing over time.

Threads the breaks through necessary methods.

Adds breaks to XHKG.

Items with TODO I'd especially like feedback on

WIP PR for updating XHKG. See details here: https://github.com/quantopian/trading_calendars/issues/152#issuecomment-689799967.

Pausing for now as:

• I need to do a bit more refactoring to get Buddha's birthday/other lunar holidays ported over correctly from pandas_market_calendars
• I need clarification from @jmccorriston on Liberation Day (I can't see that in either calendar?)
• I need some other clarification about whether we're pulling the rest of the missing holidays from PMC (which I believe to be the case, but I wanted to double check).

As the title suggests, there is always a minor failures in all calendar test tests\test_XXXX_calendar.py for all exchange code XXXX

For example, running nosetests tests\test_xnys_calendar.py yields

[success] 49.38% tests.test_xnys_calendar.XNYSCalendarTestCase.test_minute_to_session_label: 7.4921s
[success] 12.18% tests.test_xnys_calendar.XNYSCalendarTestCase.test_open_and_close_for_session: 1.8481s
[success] 11.56% tests.test_xnys_calendar.XNYSCalendarTestCase.test_next_prev_open_close: 1.7531s
[success] 11.36% tests.test_xnys_calendar.XNYSCalendarTestCase.test_next_prev_minute: 1.7233s
[success] 3.56% tests.test_xnys_calendar.XNYSCalendarTestCase.test_minute_index_to_session_labels_0: 0.5395s
[success] 3.05% tests.test_xnys_calendar.XNYSCalendarTestCase.test_next_prev_session: 0.4620s
[success] 2.26% tests.test_xnys_calendar.XNYSCalendarTestCase.test_sanity_check_session_lengths: 0.3434s
[success] 1.79% tests.test_xnys_calendar.XNYSCalendarTestCase.test_minute_index_to_session_labels_2: 0.2719s
[success] 1.78% tests.test_xnys_calendar.XNYSCalendarTestCase.test_is_open_on_minute: 0.2705s
[success] 1.73% tests.test_xnys_calendar.XNYSCalendarTestCase.test_minute_index_to_session_labels_1: 0.2623s
[success] 0.94% tests.test_xnys_calendar.XNYSCalendarTestCase.test_start_end: 0.1422s
[fail] 0.12% tests.test_xnys_calendar.XNYSCalendarTestCase.test_session_closes_in_range: 0.0187s
[success] 0.06% tests.test_xnys_calendar.XNYSCalendarTestCase.test_minutes_for_period: 0.0096s
[fail] 0.05% tests.test_xnys_calendar.XNYSCalendarTestCase.test_session_opens_in_range: 0.0069s
[success] 0.04% tests.test_xnys_calendar.XNYSCalendarTestCase.test_minutes_in_range: 0.0054s

Ran 27 tests in 15.773s

FAILED (failures=2)

Trading days fix:

• Add 2018-02-19
• Remove 2019-12-15
• Add 2020-07-10
• Add 2020-11-14

Source: https://www.tradinghours.com/exchanges/sgx/market-holidays/2018 https://www.tradinghours.com/exchanges/sgx/market-holidays/2019 https://www.tradinghours.com/exchanges/sgx/market-holidays/2020

Also changed closing time to the correct closing time from 15:00 to 17:00. This is consistent with the class' own documentation, as well as with this:

https://www.tradinghours.com/exchanges/sgx/trading-hours

Bumps actions/checkout from 2.3.4 to 3.3.0.

Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting @dependabot rebase.

Bumps actions/cache from 2.1.3 to 3.2.2.

Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting @dependabot rebase.

Bumps actions/setup-python from 2.2.1 to 4.4.0.

Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting @dependabot rebase.

TypeError Traceback (most recent call last) /var/folders/5z/nfxgl_ps6qlf5tvrct964q980000gn/T/ipykernel_864/56499229.py in ----> 1 import trading_calendars as tc

/Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages/trading_calendars/init.py in 14 # limitations under the License. 15 ---> 16 from .trading_calendar import TradingCalendar 17 from .calendar_utils import ( 18 clear_calendars,

/Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages/trading_calendars/trading_calendar.py in 31 import toolz 32 ---> 33 from .calendar_helpers import ( 34 NP_NAT, 35 compute_all_minutes,

/Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages/trading_calendars/calendar_helpers.py in 4 NANOSECONDS_PER_MINUTE = int(6e10) 5 ----> 6 NP_NAT = np.array([pd.NaT], dtype=np.int64)[0] 7 8

TypeError: int() argument must be a string, a bytes-like object or a number, not 'NaTType'