Observational Indian lunisolar calendar using the Swiss ephemeris (Hindu Drig-ganita Panchanga).

Computation of the five essentials of the panchangam:

• Tithi
• Nakshatra
• Yoga
• Karana
• Vaara

Not just the values, but also the end times of tithis and nakshatras are computed. The only factor limiting the accuracy of the program output is the uncertainity in your input values (latitude, longitude).

Also includes computation of sunrise, sunset, moonrise and moonset.

Included in the CLI version (not yet in GUI):

• Instantaneous planetary positions, including Lagna (Ascendant)
• Navamsa positions
• Choghadiya/Gauri panchanga
• Vimsottari Dasha-Bhukti
• Rahu Kala, Yamaganda Kala, Gulika Kala
• Abhijit muhurta and Durmuhurtams

By default, the month type is Amavasyanta (new moon to new moon) which is most prominent type of calendar used in South India.

NOTE: All timings are end timings. Timings displayed higher than 24:00 denote hours past midnight because the Hindu day (tithi) starts and ends with sunrise. If applicable, daylight savings (DST) are accounted for automatically based on the date and place entered in the textboxes.

Python interface to Swiss ephemeris.

     pip install pyswisseph  # OR apt-get install pyswisseph

The core of the library (panchanga.py) can be imported into other code or used from the command line.

In order to just run the GUI (gui.py) you also need python-tz and wxPython (interface to wxWidgets):

    apt-get install python-tz
    apt-get install python-wxgtk3.0

If you want to modify the GUI (Gui.wxg), you must use wxGlade:

    apt-get install python-wxglade

Wxglade 0.7.0 is buggy (0.6.8 is ok), try the development version from here.

How does it look?

Sample screenshot:

First, type the Date in DD/MM/YYYY format in the 'Date' field. Negative value for YYYY are interpolated as proleptic Gregorian calendar.

Second, type your location (city or district) in the Location field and click 'Search'. If found, then the coordinates and time zone are updated. If not, try the next method. If your location's population is more than 50,000 then the location should be found.

Third, click 'Compute'. Now the fields like tithi, etc. are computed and shown on the GUI.

First, type the Date in DD/MM/YYYY format in the 'Date' field.

Second, manually enter the coordinates and time zone of your location. You can use Google Maps or Time and Date website for this purpose.

Third, click 'Compute'. Now the fields like tithi, etc. are computed and shown on the GUI.

The program is as accurate as the Swiss Ephemeris installed on your system. So generally it is accurate for years 5000 BCE to 5000 CE, especially in the range 2500 BCE - 2500 CE. The computational speed stays the same no matter which date you enter. Compared to other software listed in the References, our software is way better in this sense.

As a simple test, try to compute the date of Madhva Navami, which is celebrated as the disappearance day of the Indian philosopher Madhvācārya. The exact date is 1317 CE, Māgha-māsa śukla-pakṣa navamī. All other software listed in References give error "Year out-of-range". But in our software, enter the place "Udipi" and date "30/1/1317" and you indeed get Māgha śuddha navamī. You can cross-verify it on the Calendrica website.

Note that dates before 1582 must be entered in proleptic Gregorian, which is a natural back-interpolation of the current Gregorian calendar we use everyday.

There are two schools of Indian calendar makers:

1. Those who follow the rules of the Sūrya Siddhāntā (SS, Theory of the Sun) or its variants like Ārya Siddhānta of Aryabhata.
2. Those who follow the Dṛk Siddhāntā (Empirical Theory).

SS contains semi-analytical equations for specifying the positions of sun and moon. However, the constants in these equations have to be updated regularly ( bīja saṃskāra ). But the equations in SS were last updated around 1000 CE, so they no longer match the planetary positions as we see today. For example, the date of solar eclipse as predicted by the equations of SS are off by many hours from its actual occurence. In spite of this, most Hindu maṭhas still publish yearly pañcāṅgas according to the rules of SS, in the name of preserving and practising tradition ( paramparā ).

The latter one, Drik school, still follow the general concepts from SS, but get the planetary positions from measured or observed data (dṛś = to see). Hence, their results match accurately with observed celestial phenomena. The Swiss Ephemeris is probably the best source available today for planetary calculations. It provides highly accurate databases of planetary data for about 10000 years. Hence, this panchanga is based on the Swiss Ephemeris. Other databases include those published by NASA's JPL (DE405) or the Moshier ephemeris.

Gaurī (Gowri) Panchanga and Choghadiya are south Indian and north Indian names respectively, for the same mathematical calculation. Basically, day and night duration are each divided into eight parts; the difference between N.Indian and S.Indian lies in their names and which part is considered auspicious/inauspicious. In the S.Indian variant, Tamilians use different order and names compared to Kannadigas/Telugus. This program provides the latter only. (This inconsitency alone is enough to let you know that such concepts of (in-)auspiciousness are all pseudo-science).

Rāhukāla, Yamagaṇḍakāla, Gulikakāla, Durmuhūrtams and Varjyam are all considered inauspicious. Abhijit muhūrta and Amṛtakāla are considered auspicious.

These planets were not discovered by Indian astronomers. They are sometimes translated as "Aruṇa graha" and "Varuṇa graha" in languages like Hindi, Nepali, etc. Problem is that there is another trans-Neptunian planet which is also called Varuna in English.

The Positional Astronomy Center translates them as हर्शल and नेपच्यून. This is inconsistent in the sense that Uranus was translated after its discoverer (William Herschel) where as Neptune was phonetically transcribed from English, instead of basing on its discoverer (Johann Galle). Therefore, I've "Indianized" their names in a rhyming fashion as हर्षल (=Uranus) and गल्ल (=Neptune). They also mean "happy" and "cheek/chin" respectively in many Indian languages.

These ones are helpful for implementing panchanga software:

• Karanam Ramakumar, Panchangam Calculations
• Second Level of the Astronomical Calculations in GCAL, used in ISKCON's GCal software.

This is the calendar book (though it mostly deals with Surya Siddhanta):

• Dershowitz and Reingold, Calendrical Calculations, 3rd edition, 2008. Online Java applet.

• Shayamasundara Dasa, Vimsottari Year -- 360 or 365 ?

Prof. M. Yanom's online interface to his Perl code -- this is the best version of the old Surya Siddhanta pancanga I've seen. However, the Surya Siddhanta system (no fault with the Perl code) is not accurate if you want to work with dates which are several centuries before our current time.

drikpanchang is a reliable online calendar for the Drik. However, it is neither open source nor do they have a desktop program. This website doesn't work for dates before 1600 CE. Their Android app doesn't work for dates outside the range 1900 - 2100 CE.

Hindu Calendar for Android is another offline Drik calendar by Alok Mandavgane. Again, this is not open source. This software has a bug that it doesn't account for daylight savings in Europe. Also it doesn't work for dates outside the range 1900 - 2100 CE.

Among open source programs, I found these two:

• On Google Code: generates a pdf of panchanga for any year and place, but imprecise. For ex., tithi end timings are off by ten minutes sometimes. There is no GUI either.

• On GitHub: Based on Paul Schlyter's semi-analytical model for planetary positions. This program gives the panchanga for a given instant but doesn't ask for a place's coordinates or timezone. This is probably because the program doesn't compute sunrise timings at all! The planetary model fails for dates outside the range 1800 CE to 2100 CE.

GNU Affero GPL version 3 (or later).

• Festivals
• Amritakala, varjyam
• gettext translations
• Harmonize all functions to use UT aka UT1 instead of UTC or ET (swe.jdut1_to_utc() <==> swe.utc_to_jd()[1], swe.utc_time_zone(), etc.)