dayfilter 0.3.1

Filter time-series based on sun movement

dayfilter

Filter time-series based on sun movement

Installation

pip install dayfilter

DayFilter depends on suntime. Due to known issue of suntime, please use my fork of suntime on test-pypi. To install, use:

pip uninstall -y suntime
pip install -i https://test.pypi.org/simple/ suntime-yasirroni

I can't bundle my fork of suntime into DayFilter due to LICENSE incompatibility issue.

Usage

Simple use case of getting list of bool

from datetime import datetime
from dayfilter import is_daytime

date_string = '2010-01-01-16:58:30.00000'
date_datetime = datetime.strptime(date_string, '%Y-%m-%d-%H:%M:%S.%f')

# WASHINGTON DC DULLES INT'L AR [STERLING - ISIS], VA. data
latitude = 38.98
longitude = -77.47
time_zone = -5

is_daytime(date_datetime, latitude, longitude, time_zone)

Using DayFilter class for multiple call

from dayfilter import DayFilter
import pandas as pd
from datetime import datetime

# WASHINGTON DC DULLES INT'L AR [STERLING - ISIS], VA. data
latitude = 38.98
longitude = -77.47
time_zone = -5

f = DayFilter(latitude, longitude, time_zone)

date_strings = [
    '2010-01-01-05:58:42.00000',
    '2010-01-01-06:59:22.00000',
    '2010-01-01-13:22:23.00000',
    '2010-01-01-14:54:55.00000',
    '2010-01-01-15:31:45.00000',
    '2010-01-02-17:12:22.00000',
    '2010-01-02-15:12:32.00000',
    '2010-01-02-16:59:32.00000',
    '2010-01-02-18:00:11.00000',
]
date_format = '%Y-%m-%d-%H:%M:%S.%f'
date_datetimes = [datetime.strptime(date_string, date_format) for date_string in date_strings]

df = pd.DataFrame(index=date_datetimes, data=[1,2,3,4,5,6,7,8,9])

df_ = f.filter(df)
print(df_)

Adcanced use of DayFilter class with custom filter

from dayfilter import DayFilter
from dayfilter.post_process import shift_sr_down, shift_ss_up
from dayfilter.logic import logic_daytime

f2 = DayFilter(latitude, longitude, time_zone, post_processes=[shift_sr_down, shift_ss_up], logic=logic_daytime)

df__ = f2.filter(df)
print(df__)

print(f.get_sr_ss(df_.index[0]))
print(f.get_sr_ss(df_.index[0], post_processes=True))

TODO

1. Consider refactor to astral due to suntime LICENSE. a. Astral is slower and have same bug with the original suntime b. Soulution of the suntime bug exist on yasirroni/suntime

License

This project is licensed in MIT License, but depends on suntime that use LGPL license.