Code for parsing lat-long coordinates in "various" formats, and for converting between lat-long formats (e.g. decimal degrees to degrees-minutes-seconds)
The goal is to abide by Postel's Law: it will accept almost anything that can be unambiguously interpreted a latitude or longitude coordinates. It's not looking for particular symbols, etc.
In [12]: from lat_lon_parser import parse In [13]: parse("45° 12.6' W") Out[13]: -45.21
Decimal degrees (easy):
23.43 -45.21
Decimal Degrees with hemisphere:
23.43 N 45.21 W
Or with spelled out:
23.43 North 45.21 West
(note that all of the cardinal directions are not case-sensitive)
Degrees, decimal minutes: (now it starts getting tricky!):
23° 25.800' -45° 12.600'
or:
23 25.800' -45 12.600'
or:
23° 25.8' N 45° 12.6' West
Degrees, Minutes, Seconds: (really fun!!!):
23° 25' 48.0" -45° 12' 36.0"
or:
23d 25' 48.0" -45d 12' 36.0"
or:
23° 25' 48.0" North 45° 12' 36.0" S
or -- lots of other combinations!
For a more complete list, see the tests
This uses a pretty "stupid" algorithm -- it assumes that all formats will be something like:
[-][space] degrees [separator] minutes [separator] seconds [separator] [N[orth]|S[outh|E[ast]|W[est]]
But that actually is pretty darn robust!
If you have other formats you want to be able to parse, please contribute tests! -- And ideally a patch if the current code doesn't work.
Also included is code to convert to other formats used for latitude and longitude:
- degrees
- degrees minutes
- degrees minutes seconds
Functions for returning tuples of numbers:
>>> to_dec_deg(23, 12, 3) 23.200833333333332 >>> to_deg_min(34.1234) (34.0, 7.404) >>> to_deg_min_sec(34.1234) (34.0, 7, 24.24)
Functions for converting to various string formats:
>>> to_str_dec_deg(23, 12, 3) '23.200833°' >>> to_str_deg_min(2.345) "2° 20.700'" >>> to_str_deg_min_sec(-23.1234) '-23° 7\' 24.24"' >>> to_str(23.45) '23.450000°' >>> to_str(23, 45) "23° 45.000'" >>> to_str(23, 45, 6.7) '23° 45\' 6.70"'
Question from a user:
How to apply this lat_lon_parser on pandas dataframe specific column?
Turns out it's straightforward -- just pass the parse function to apply:
In [20]: df = pandas.DataFrame({'coords':["12d13'N","32 5 14", "30.123W"]}) In [21]: df Out[21]: coords 0 12d13'N 1 32 5 14 2 30.123W In [22]: df['coords'] = df['coords'].apply(parse) In [23]: df Out[23]: coords 0 12.216667 1 32.087222 2 -30.123000