business-python 1.0.0

Date calculations based on business calendars.

Business (Python)

Date calculations based on business calendars. (Python 3.6+)

Python implementation of https://github.com/gocardless/business

Documentation

To get business, simply:

$ pip install business-python

Getting started

Get started with business by creating an instance of the calendar class, passing in a hash that specifies which days of the week are considered working days, and which days are holidays.

from business.calendar import Calendar

calendar = Calendar(
  working_days=["monday", "tuesday", "wednesday", "thursday", "friday"],
  # array items are either parseable date strings, or real datetime.date objects
  holidays=["January 1st, 2020", "April 10th, 2020"],
  extra_working_dates=[],
)

extra_working_dates key makes the calendar to consider a weekend day as a working day.

A few calendar configs are bundled with the package (see business/data for details). Load them by calling the load class method on Calendar.

calendar = Calendar.load("weekdays")

If working_days is missing, then common default is used (mon-fri). If holidays is missing, "no holidays" assumed. If extra_working_dates is missing, then no changes in working_days will happen.

Elements of holidays and extra_working_dates may be either strings that Calendar.parse_date() can understand, or YYYY-MM-DD (which is considered as a Date by Python YAML itself).

holidays:
  - 2017-01-08 # Same as January 8th, 2017

The load_cache method allows a thread safe way to avoid reloading the same calendar multiple times, and provides a performant way to dynamically load calendars for different requests.

calendar = Calendar.load_cache("weekdays")

Input data types

The parse_date method is used to process the input date(s) in each method and return a datetime.date object.

Calendar.parse_date("2019-01-01")
# => datetime.date(2019, 1, 1)

Supported data types are:

• datetime.date
• datetime.datetime
• pandas.Timestamp (treated as datetime.datetime)
• date string parseable by dateutil.parser.parse

numpy.datetime64 is not supported, but can be converted to datetime.date:

numpy.datetime64('2014-06-01T23:00:05.453000000').astype('M8[D]').astype('O')
# =>  datetime.date(2014, 6, 1)

Checking for business days

To check whether a given date is a business day (falls on one of the specified working days or extra working dates, and is not a holiday), use the is_business_day method on Calendar.

calendar.is_business_day("Monday, 8 June 2020")
# => true
calendar.is_business_day("Sunday, 7 June 2020")
# => false

Custom calendars

To use a calendar you've written yourself, you need to add the directory it's stored in as an additional calendar load path:

Calendar.additional_load_paths = ['path/to/your/calendar/directory']

You can then load the calendar as normal.

Business day arithmetic

For our purposes, date-based calculations are sufficient. Supporting time-based calculations as well makes the code significantly more complex. We chose to avoid this extra complexity by sticking solely to date-based mathematics.

The add_business_days method is used to perform business day arithmetic on dates.

input_date = Calendar.parse_date("Thursday, 12 June 2014")
calendar.add_business_days(input_date, 4).strftime("%A, %d %B %Y")
# => "Wednesday, 18 June 2014"
calendar.add_business_days(input_date, -4).strftime("%A, %d %B %Y")
# => "Friday, 06 June 2014"

The roll_forward and roll_backward methods snap a date to a nearby business day. If provided with a business day, they will return that date. Otherwise, they will advance (forward for roll_forward and backward for roll_backward) until a business day is found.

input_date = Calendar.parse_date("Saturday, 14 June 2014")
calendar.roll_forward(input_date).strftime("%A, %d %B %Y")
# => "Monday, 16 June 2014"
calendar.roll_backward(input_date).strftime("%A, %d %B %Y")
# => "Friday, 13 June 2014"

In contrast, the next_business_day and previous_business_day methods will always move to a next or previous date until a business day is found, regardless if the input provided is a business day.

input_date = Calendar.parse_date("Monday, 9 June 2014")
calendar.roll_forward(input_date).strftime("%A, %d %B %Y")
# => "Monday, 09 June 2014"
calendar.next_business_day(input_date).strftime("%A, %d %B %Y")
# => "Tuesday, 10 June 2014"
calendar.previous_business_day(input_date).strftime("%A, %d %B %Y")
# => "Friday, 06 June 2014"

To count the number of business days between two dates, pass the dates to business_days_between. This method counts from start of the first date to start of the second date. So, assuming no holidays, there would be two business days between a Monday and a Wednesday.

from datetime import timedelta

input_date = Calendar.parse_date("Saturday, 14 June 2014")
calendar.business_days_between(input_date, input_date + timedelta(days=7))
# => 5

The get_business_day_of_month method return the running total of business days for a given date in that month. This method counts the number of business days from the start of the first day of the month to the given input date.

input_date = Calendar.parse_date("Thursday, 12 June 2014")
calendar.get_business_day_of_month(input_date)
# => 9

Included Calendars

We include some calendar data with this package but give no guarantees of its accuracy. The calendars that we include are:

• ACH (United States)
• Bacs
• Bankgirot
• BECS (Australia)
• BECSNZ (New Zealand)
• PAD (Canada)
• Betalingsservice
• Target (SEPA)
• TargetFrance (SEPA + French bank holidays)

License & Contributing

• This is available as open source under the terms of the MIT License.
• Bug reports and pull requests are welcome on GitHub at https://github.com/gocardless/business-python.

GoCardless ♥ open source. If you do too, come join us.