def holidays(self):
import holidays
country_code = self.iso.split('-')[0]
if self.region_type == self.STATE:
state_code = self.iso.split('-')[1]
return holidays.CountryHoliday(country_code, state=state_code)
elif self.region_type == self.PROVINCE:
prov_code = self.iso.split('-')[1]
return holidays.CountryHoliday(country_code, prov=prov_code)
return holidays.CountryHoliday(country_code)
def holiday_flag(DTindex, country: str = 'US'):
"""Create a 0/1 flag for given datetime index.
Args:
DTindex (panda.DatetimeIndex): DatetimeIndex of dates to create flags
country (str): to pass through to python package Holidays
Returns:
pandas.Series() with DatetimeIndex and column 'HolidayFlag'
"""
if country.upper() == 'US':
try:
import holidays
country_holidays = holidays.CountryHoliday('US')
country_holidays = country_holidays[DTindex[0]:DTindex[-1]]
all_days = pd.Series(np.repeat(0, len(DTindex)), index=DTindex)
holi_days = pd.Series(
np.repeat(1, len(country_holidays)),
index=pd.DatetimeIndex(country_holidays),
)
holi_days = all_days.combine(holi_days, func=max).fillna(0)
holi_days.rename("HolidayFlag", inplace=True)
except Exception:
from pandas.tseries.holiday import USFederalHolidayCalendar
# uses pandas calendar as backup in the event holidays fails
holi_days = (USFederalHolidayCalendar().holidays().to_series()
[DTindex[0]:DTindex[-1]])
all_days = pd.Series(np.repeat(0, len(DTindex)), index=DTindex)
holi_days = pd.Series(np.repeat(1, len(holi_days)),
index=holi_days)
holi_days = all_days.combine(holi_days, func=max).fillna(0)
holi_days.rename("HolidayFlag", inplace=True)
else:
import holidays
country_holidays = holidays.CountryHoliday(country.upper())
country_holidays = country_holidays[DTindex[0]:DTindex[-1]]
all_days = pd.Series(np.repeat(0, len(DTindex)), index=DTindex)
holi_days = pd.Series(
np.repeat(1, len(country_holidays)),
index=pd.DatetimeIndex(country_holidays),
)
holi_days = all_days.combine(holi_days, func=max).fillna(0)
holi_days.rename("HolidayFlag", inplace=True)
return holi_days[DTindex]
def get_holidays(tickers_df, ticker_name):
country_to_code = {
'Netherlands': 'NLD',
'France': 'FRA',
'Luxembourg': 'LUX',
'Switzerland': 'CHE',
'United Kingdom': 'GBR',
'Ireland': 'IE',
'Russian Federation': 'RUS',
'Mexico': 'MEX',
'Bermuda': 'USA',
'United States': 'USA',
'Germany': 'DE',
'China': 'USA',
'Israel': 'ISR'
}
holidays_df = holidays.CountryHoliday(
country_to_code[tickers_df.loc[ticker_name]['country']], years=YEARS)
holidays_df = pd.DataFrame.from_dict(holidays_df, orient="index")
holidays_df["ds"] = holidays_df.index
holidays_df["holiday"] = holidays_df[0]
del holidays_df[0]
return holidays_df
def is_holiday(date, country_code):
# Requirement: holidays python library (https://pypi.org/project/holidays/)
# Install: pip install holidays
try:
import holidays
isCountryFound = False
for country in holidays.list_supported_countries():
if country_code == country:
print(('Found country code "{}"'.format(country_code)))
country_holidays = holidays.CountryHoliday(country_code)
isCountryFound = True
break
if (isCountryFound == False):
print((
'Error: given country code "{}" is not found in supported country codes'
.format(country_code)))
print(holidays.list_supported_countries())
return False
return date.date() in country_holidays
except ImportError as err:
print('*** Could not import "holidays" Python module ***')
print(
'Please install "holidays", or Holidays are not taken into account.'
)
print('To install, type pip install holidays')
return False
def adjust_holiday(trade_file):
country_list = list(trade_file["listing_place"].drop_duplicates())
start_date_dict = defaultdict(list)
end_date_dict = defaultdict(list)
for country in country_list:
this_country_df = trade_file[trade_file["listing_place"] == country].copy()
this_holiday = holidays.CountryHoliday(country)
holiday_start_date = [start_date for start_date in list(this_country_df["trade_start_date"].drop_duplicates())
if start_date in this_holiday]
holiday_end_date = [end_date for end_date in list(this_country_df["trade_end_date"].drop_duplicates())
if end_date in this_holiday]
if holiday_start_date != []:
for date in holiday_start_date:
adjust_date = date + BDay(1)
while adjust_date in this_holiday:
adjust_date += BDay(1)
start_date_dict[date] = [adjust_date]
if holiday_end_date != []:
for date in holiday_end_date:
adjust_date = date - BDay(1)
while adjust_date in this_holiday:
adjust_date -= BDay(1)
end_date_dict[date] = [adjust_date]
adjust_num = 0
for adjust_dict in [start_date_dict, end_date_dict]:
date_column = "trade_start_date" if adjust_num == 0 else "trade_end_date"
for key, value in adjust_dict.items():
adjust_index = trade_file[trade_file[date_column] == key].index
trade_file.loc[adjust_index, date_column] = value[0]
adjust_num += 1
return trade_file
def transform(self, x: DataFrame, y: Series = None) -> DataFrame:
x.loc[:, "measure_date"] = pd.to_datetime(
pd.to_datetime(x["date_time"]).dt.date)
x.loc[:, "departure_date"] = pd.to_datetime(
pd.to_datetime(x["departure_date_time"]).dt.date)
x.loc[:, "departure_hour"] = pd.to_datetime(
x["departure_date_time"]).dt.hour
x.loc[:, "departure_minute"] = pd.to_datetime(
x["departure_date_time"]).dt.minute
x.loc[:, "arrival_date"] = pd.to_datetime(
pd.to_datetime(x["arrival_date_time"]).dt.date)
x.loc[:,
"arrival_hour"] = pd.to_datetime(x["arrival_date_time"]).dt.hour
x.loc[:, "arrival_minute"] = pd.to_datetime(
x["arrival_date_time"]).dt.minute
ua_holidays = []
for date in holidays.CountryHoliday("UA", prov=None,
years=[2020]).items():
ua_holidays.append(str(date[0]))
x.loc[:, "is_holiday"] = [
1 if str(val).split()[0] in ua_holidays else 0
for val in x["departure_date"]
]
x.loc[:, "days_left_to_departure"] = (
x["departure_date"] -
x["measure_date"]).transform(lambda x: x.dt.days)
del x["departure_date_time"]
del x["arrival_date_time"]
return x
def get_holidays(start='2015-01-01', country='DK', frequency='D'):
"""
Takes in a start date and a country.
Produces a dataframe with a daily date time index and columns:
day_of_week - numerical day of the week identifier 0 for monday
holiday_bool - boolean true or false for holiday
holiday_name - name of the holiday if holiday_bool is true
Returns a dataframe
"""
# get end date
end = str(date.today() + timedelta(3))
#generate the range of daily dates
dates = pd.date_range(start=start, end=end, freq=frequency)
#create the holiday object
country_holidays = holidays.CountryHoliday(country)
#create a list for the holiday bool and name
holiday_list = []
#loop through the dates
for d in dates:
#true if holiday in object, false otherwise
holiday_bool = d in country_holidays
holiday_names = country_holidays.get(d)
holiday_list.append([holiday_bool, holiday_names])
#create return dataframe
holidays_data = pd.DataFrame(holiday_list,
index=dates,
columns=['holiday', 'holiday_name'])
holidays_data.holiday = holidays_data.holiday.astype('int')
# add whether it is weekend
holidays_data['weekend'] = 0
holidays_data.loc[(holidays_data.index.dayofweek == 5) |
(holidays_data.index.dayofweek == 6), 'weekend'] = 1
return holidays_data
def next_business_day(self, date, time_span, time_unit, country_code):
ONE_DAY = datetime.timedelta(days=1)
HOLIDAYS = holidays.CountryHoliday(country_code)
next_day = date + ONE_DAY
while next_day.weekday() in holidays.WEEKEND or next_day in HOLIDAYS:
next_day += ONE_DAY
return next_day
def __init__(self, start_date, end_date, user_id, token):
self.start_date = datetime.strptime(start_date, "%Y-%m-%d").date()
self.end_date = datetime.strptime(end_date, "%Y-%m-%d").date()
self.user_id = user_id
self.token = token
self.holidays = holidays.CountryHoliday("SE", False)
print("Start date: %s, end date: %s, user ID: %s" % (self.start_date, self.end_date, self.user_id))
def holidays_timeseries(time_index,
country_code: str,
prov: str = None,
state: str = None) -> TimeSeries:
"""
Creates a binary univariate TimeSeries with index `time_index` that equals 1 at every index that lies within
(or equals) a selected country's holiday, and 0 otherwise.
Available countries can be found `here <https://github.com/dr-prodigy/python-holidays#available-countries>`_.
Parameters
----------
country_code
The country ISO code
prov
The province
state
The state
Returns
-------
TimeSeries
A new binary holiday TimeSeries instance.
"""
country_holidays = holidays.CountryHoliday(country_code,
prov=prov,
state=state)
scoped_country_holidays = country_holidays[time_index[0]:time_index[-1] +
pd.Timedelta(days=1)]
index_series = pd.Series(time_index, index=time_index)
values = index_series.apply(lambda x: x in scoped_country_holidays).astype(
int)
return TimeSeries.from_times_and_values(time_index, values)
def add_time_information(
inpath=os.path.join("data", "interim", "data_indexed_converted.csv"),
outpath=os.path.join("data", "interim",
"data_indexed_converted_timed.csv"),
):
"""Adds columns for year, month, weekday, hour and holiday
Keyword Arguments:
inpath {path} -- [path for incoming data] (default: {os.path.join("data", "interim", "data_indexed_converted.csv")})
outpath {path} -- [path for where to save data] (default: {os.path.join("data", "interim", "data_indexed_converted_timed.csv")})
"""
import holidays
fra_holidays = holidays.CountryHoliday("FRA")
data = pd.read_csv(inpath,
parse_dates=["Date_Time"],
index_col="Date_Time")
data["year"] = data.index.year
data["month"] = data.index.month_name()
data["weekday"] = data.index.weekday_name
data["hour"] = data.index.hour
data["date"] = data.index.date
data["holiday"] = data["date"].apply(lambda x: x in fra_holidays)
data.drop(columns=["date"], inplace=True)
data.to_csv(outpath)
def __init__(self,
workweek_schedule=_DEFAULT_WORKWEEK_SCHEDULE,
lunch_hour=12,
offdays="US"):
"""Keyword arguments:
workweek_schedule --
a dict that defines weekly work hours; has following format
{
0: (9, 17),
...
6: ()
}
The key is an int following datetime weekday() function where
Monday is 0 and Sunday is 6. The value is a tuple with two ints:
the start hour and the end hour in 24 hour format (0 to 23).
lunch_hour -- int
offdays -- either an ISO country code or tuple of datetime dates to be
considered as holidays
"""
_ok, _errcode = _verify_init(workweek_schedule, lunch_hour, offdays)
if not _ok:
raise ValueError(_ERRCODE_MESSAGE[_errcode])
if isinstance(offdays, str):
self._offdays = holidays.CountryHoliday(offdays)
else:
self._offdays = offdays
self._workweek_schedule = workweek_schedule
self._lunch_hour = lunch_hour
self._workhour_lookup = _build_workhour_lookup(workweek_schedule,
lunch_hour)
def get_items(country, year, subdivision_type, subdivisions):
items = []
for subdivision in subdivisions:
state = None
prov = None
if subdivision_type == "state":
state = subdivision
prov = None
elif subdivision_type == "prov":
state = None
prov = subdivision
for ptr in holidays.CountryHoliday(country,
state=state,
prov=prov,
years=year,
expand=False).items():
if (subdivision_type == "none"):
location = country
else:
location = '{0}-{1}'.format(country, subdivision)
items.append({
"region": location,
"name": ptr[1],
"date": ptr[0].isoformat(),
"req_year": year
})
return items
def get_default_holidays(times, country):
"""Creates default holidays for a specific country.
Args:
times: a Pandas `DatetimeIndex` that indexes time series data.
country: `str`, two-letter upper-case [ISO 3166-1 alpha-2 country code](
https://en.wikipedia.org/wiki/ISO_3166-1_alpha-2) for using holiday
regressors from a particular country.
Returns:
holidays: a Pandas `DataFrame` with default holidays relevant to the input
times. The `DataFrame` should have the following columns:
* `geo`: `str`, two-letter upper-case country code
* `holiday`: `str`, the name of the holiday
* `date`: `str`, dates in the form of `YYYY-MM-DD`
"""
# pylint: disable=g-import-not-at-top
import pandas as pd
import holidays
# pylint: enable=g-import-not-at-top
years = range(times.min().year, times.max().year + 1)
holidays = holidays.CountryHoliday(country, years=years, expand=False)
holidays = pd.DataFrame(
[(country, holidays.get_list(date), date) for date in holidays],
columns=['geo', 'holiday', 'date'])
holidays = holidays.explode('holiday')
# Ensure that only holiday dates covered by times are used.
holidays = holidays[(holidays['date'] >= times.min())
& (holidays['date'] <= times.max())]
holidays = holidays.reset_index(drop=True)
holidays['date'] = pd.to_datetime(holidays['date'])
holidays = holidays.sort_values('date')
holidays['date'] = holidays['date'].dt.strftime('%Y-%m-%d')
return holidays
def flag_holiday(day, country, prov=None):
"""Return a flag if the day given is a holiday according to country, province and years in the
format required by the holidays package.
day: datetime object with with the day that needs to be queried.
country: ISO 3166 code for the country.
prov: ISO 3166-2 code for the province/state
>>> import datetime
>>> day1 = datetime.datetime.strptime('2017-12-26', '%Y-%m-%d')
>>> flag_holiday(day1, 'CA', 'ON')
1
>>> day2 = datetime.datetime.strptime('2019-07-04', '%Y-%m-%d')
>>> flag_holiday(day2, 'US')
1
"""
import holidays
#Get datetimes
holidays_list = list(holidays.CountryHoliday(country=country, years=day.year, prov=prov).keys())
# Format as strings
holidays_dates = [day.strftime('%Y-%m-%d') for day in holidays_list]
# Compare
flag = day.strftime('%Y-%m-%d') in holidays_dates
return int(flag)
def is_business_day(date, nation):
if nation == "FR":
nation = "FRA"
try:
return date in holidays.CountryHoliday(nation)
except KeyError:
return False
def holidays_list() -> list:
"""Return list of holidays for current and upcoming year."""
year = datetime.today().year
ca_holidays = holidays.CountryHoliday("US",
prov=None,
state="CA",
years=[year, year + 1])
return [i.isoformat() for i in ca_holidays]
def get_holiday(start_year, end_year, country = 'us'):
output_date_list = []
us_holiday = holidays.CountryHoliday('US')
year_len = end_year - start_year + 1
for _ in range(year_len):
output_date_list.extend(us_holiday[f'{start_year}-01-01':f'{start_year}-12-31'])
start_year += 1
return output_date_list
def set_holidays(df, country='US', state='CA', prov=None):
"""Flag holidays received on weekends."""
holiday_list = holidays.CountryHoliday(country, prov=prov, state=state)
is_holiday_func = lambda dttm: 1 if dttm in holiday_list else 0
df['is_holiday'] = df['Received DtTm'].apply(is_holiday_func)
return df
def date_is_holiday(array):
"""
is_holiday([ pd.to_datetime("2015/1/1") ] * 10)
"""
import holidays , numpy as np
jp_holidays = holidays.CountryHoliday('JP')
return np.array( [ 1 if x.astype('M8[D]').astype('O') in jp_holidays else 0 for x in array] )
def get_observed_holiday_name(year, month, day):
name = holidays.CountryHoliday('US', state='TX').get(
datetime.date(year, month, day))
if name and name.replace('(Observed)',
'').strip() in Holiday.__OBSERVED_HOLIDAYS:
return name
else:
return None
def working_day(dt: str | datetime,
country: str = 'BR',
state: str | None = None) -> bool:
"""
Indices if a day specified by the user is a working day.
Parameters
----------
dt : str or datetime
Specifies the day the user wants to know if it is a business day.
country : str
Indicates country to check for vacation days, by default 'BR'
state: str
Indicates state to check for vacation days, by default None
Returns
-------
boolean
if true, means that the day informed by the user is a working day.
if false, means that the day is not a working day.
Examples
--------
>>> from pymove.utils.datetime import str_to_datetime
>>> independence_day = str_to_datetime('2021-09-7 12:00:01') # Holiday in Brazil
>>> next_day = str_to_datetime('2021-09-8 12:00:01') # Not a Holiday in Brazil
>>> print(working_day(independence_day, 'BR'))
False
>>> print(type(working_day(independence_day, 'BR')))
<class 'bool'>
>>> print(working_day(next_day, 'BR'))
True
>>> print(type(working_day(next_day, 'BR')))
'<class 'bool'>'
References
----------
Countries and States names available in https://pypi.org/project/holidays/
"""
result = True
if isinstance(dt, str):
dt = str_to_datetime(dt)
if isinstance(dt, datetime):
dt = datetime(dt.year, dt.month, dt.day)
if dt in holidays.CountryHoliday(country=country, prov=None, state=state):
result = False
else:
dow = to_day_of_week_int(dt)
# 5 == Saturday, 6 == Sunday
if dow == 5 or dow == 6:
result = False
return result
def hdays(df,column_name):
holiday_date=[]
holis = holidays.CountryHoliday('IND')
for i,j in holis.items():
holiday_date.append(i)
list_holidays=Series([holis.get(i) for i in df[column_name]])
list_holidays=Series([i if i !=None else 'Normal' for i in list_holidays])
df['holi']=list_holidays
return df
def is_holiday(self, hdate):
now = datetime.now()
#print(now.year)
nic_holidays = holidays.CountryHoliday(country='NI', years=now.year)
print(hdate)
#print(nic_holidays)
print("IS HOLIDAY?")
print(hdate in nic_holidays)
return hdate in nic_holidays
def _holidays(min_date, max_date, countries=None):
dates = pd.date_range(min_date, max_date, freq="D")
d = {}
if countries is None:
countries = ['FRA', 'AU', 'US', 'DE', 'custom']
for c in countries:
if c != 'custom':
h = holidays.CountryHoliday(c)
col_name = "is_holiday_{}".format(c.lower())
d[col_name] = dates.map(lambda x: x in h)
else:
custom_countries = [
'US', "PL", "AU", "FRA", "DE", "CZ", "SK", "ES", "PT",
"CH", "NZ", "UK", "IT", "AT", "BE", "ZA"
]
h_set = [holidays.CountryHoliday(c) for c in custom_countries]
col_name = "is_holiday_custom"
d[col_name] = dates.map(lambda x: any([x in h for h in h_set]))
return pd.DataFrame(d, index=dates)
def holidays_timeseries(
time_index: pd.DatetimeIndex,
country_code: str,
prov: str = None,
state: str = None,
column_name: Optional[str] = "holidays",
until: Optional[Union[int, str, pd.Timestamp]] = None,
add_length: int = 0,
dtype: np.dtype = np.float64,
) -> TimeSeries:
"""
Creates a binary univariate TimeSeries with index `time_index` that equals 1 at every index that lies within
(or equals) a selected country's holiday, and 0 otherwise.
Available countries can be found `here <https://github.com/dr-prodigy/python-holidays#available-countries>`_.
Parameters
----------
time_index
The time index over which to generate the holidays
country_code
The country ISO code
prov
The province
state
The state
until
Extend the time_index up until timestamp for datetime indexed series
and int for range indexed series, should match or exceed forecasting window.
add_length
Extend the time_index by add_length, should match or exceed forecasting window.
Set only one of until and add_length.
column_name
Optionally, the name of the value column for the returned TimeSeries
dtype
The desired NumPy dtype (np.float32 or np.float64) for the resulting series
Returns
-------
TimeSeries
A new binary holiday TimeSeries instance.
"""
time_index = _extend_time_index_until(time_index, until, add_length)
scope = range(time_index[0].year,
(time_index[-1] + pd.Timedelta(days=1)).year)
country_holidays = holidays.CountryHoliday(country_code,
prov=prov,
state=state,
years=scope)
index_series = pd.Series(time_index, index=time_index)
values = index_series.apply(lambda x: x in country_holidays).astype(dtype)
return TimeSeries.from_times_and_values(time_index,
values,
columns=pd.Index([column_name]))
def ip_localisation(ip, user):
response = DbIpCity.get("92.184.112.171", api_key='free')
get_holidays = holidays.CountryHoliday(response.country)
today = datetime.datetime.today()
is_holiday = today.strftime("%Y-%m-%d") in get_holidays
Login.objects.create(user=user,
location=response.country,
is_holiday=is_holiday)
def get_is_holiday_from_series(series: pd.Series,
country: str = "UnitedKingdom") -> pd.Series:
"""Return 1 if day is a public holiday.
By default, uses UK holidays, but can specify a country by string name in
`country` arg. See `holidays.list_supported_countries()` for list of
supported countries.
"""
years = series.dt.year.unique()
holiday_dates = holidays.CountryHoliday(country, years=years)
return series.dt.date.isin(holiday_dates).astype(int)
def __init__(self, hass, config):
self.config = config
self.__name = config.get(CONF_NAME)
self.__frequency = config.get(CONF_FREQUENCY)
self.__collection_days = config.get(CONF_COLLECTION_DAYS)
first_month = config.get(CONF_FIRST_MONTH)
if first_month in MONTH_OPTIONS:
self.__first_month = MONTH_OPTIONS.index(first_month) + 1
else:
self.__first_month = 1
last_month = config.get(CONF_LAST_MONTH)
if last_month in MONTH_OPTIONS:
self.__last_month = MONTH_OPTIONS.index(last_month) + 1
else:
self.__last_month = 12
self._weekday_order_numbers = config.get(CONF_WEEKDAY_ORDER_NUMBER)
self._week_order_numbers = config.get(CONF_WEEK_ORDER_NUMBER)
self.__monthly_force_week_numbers = bool(
self._week_order_numbers is not None and len(self._week_order_numbers) != 0
)
self.__include_dates = to_dates(config.get(CONF_INCLUDE_DATES, []))
self.__exclude_dates = to_dates(config.get(CONF_EXCLUDE_DATES, []))
country_holidays = config.get(CONF_MOVE_COUNTRY_HOLIDAYS)
self.__holidays = []
if country_holidays is not None and country_holidays != "":
today = dt_util.now().date()
this_year = today.year
years = [this_year, this_year + 1]
try:
for date, name in holidays.CountryHoliday(
country_holidays, years=years
).items():
if date >= today:
self.__holidays.append(date)
except KeyError:
_LOGGER.error("Invalid country code (%s)", country_holidays)
_LOGGER.debug("(%s) Found these holidays %s", self.__name, self.__holidays)
self.__period = config.get(CONF_PERIOD)
self.__first_week = config.get(CONF_FIRST_WEEK)
self.__first_date = to_date(config.get(CONF_FIRST_DATE))
self.__next_date = None
self.__today = None
self.__days = 0
self.__date = config.get(CONF_DATE)
self.__entities = config.get(CONF_ENTITIES)
self.__verbose_state = config.get(CONF_VERBOSE_STATE)
self.__state = "" if bool(self.__verbose_state) else 2
self.__icon_normal = config.get(CONF_ICON_NORMAL)
self.__icon_today = config.get(CONF_ICON_TODAY)
self.__icon_tomorrow = config.get(CONF_ICON_TOMORROW)
self.__date_format = config.get(CONF_DATE_FORMAT, DEFAULT_DATE_FORMAT)
self.__verbose_format = config.get(CONF_VERBOSE_FORMAT, DEFAULT_VERBOSE_FORMAT)
self.__icon = self.__icon_normal
def is_work_day(date_time: datetime) -> bool:
"""
Determine the if the date_time is a business day in BC.
Assumes Sat, Sun and all BC stat holidays are non-working days.
Requires 3rd party library: holidays==0.10.4 or better
"""
day_of_week = date_time.strftime("%a")
if not (day_of_week == "Sun" or day_of_week == "Sat"):
if date_time not in holidays.CountryHoliday('CA', prov='BC'):
return True
return False