import os
import datetime
import pandas as pd
import numpy as np
from pyspedas.utilities.download import download
from pyspedas.tplot_tools import time_double
[docs]
def load_leap_table(reload=False):
"""
Loads the leap second table for converting TAI to unix times
Parameters
-----------
reload: bool
Re-load the leap second table, even if it exists locally.
This shouldn't be needed until at least 2035:
https://www.scientificamerican.com/article/the-leap-seconds-time-is-up-world-votes-to-stop-pausing-clocks/
Returns
---------
dict
A dictionary containing 'dates' with array of Julian dates corresponding to the leap seconds in the 'leaps' array
"""
if os.environ.get('CDF_LEAPSECONDSTABLE') is not None:
table_file = os.environ.get('CDF_LEAPSECONDSTABLE')
elif os.environ.get('SPEDAS_DATA_DIR') is not None:
table_file = os.path.join(os.environ.get('SPEDAS_DATA_DIR'), 'CDFLeapSeconds.txt')
else:
table_file = os.path.join('data', 'CDFLeapSeconds.txt')
table_dir = os.path.dirname(table_file)
if reload or not os.path.exists(table_file):
downloaded = download(remote_path='https://cdf.gsfc.nasa.gov/html/',
remote_file='CDFLeapSeconds.txt',
local_path=table_dir)
cols = ['Year', 'Month', 'Day', 'LS', 'Drift1','Drift2']
table = pd.read_csv(table_file,
sep=r'\s+',
dtype=str,
names=cols,
comment=';',
skipinitialspace=True,
index_col=False)
leap_dates = table['Year'].to_numpy() + '-' + table['Month'].to_numpy() + '-' + table['Day'].to_numpy()
leap_dates = time_double(leap_dates)
juls = np.array(leap_dates)/86400.0 + datetime.date(1970, 1, 1).toordinal() + 1721424.5
return {'leaps': np.float64(table['LS'].to_numpy()),
'juls': juls}