2014-11-17 00:10:43 +00:00
|
|
|
"""
|
|
|
|
thesethreewords: Human-readable addresses for every 3x3m square of the earth's surface.
|
|
|
|
The simplest ways to use this module are the `three_words` and `decode`
|
|
|
|
functions. For more see `WordHasher`.
|
|
|
|
"""
|
2014-11-17 00:03:10 +00:00
|
|
|
import random
|
|
|
|
|
|
|
|
import geohash
|
|
|
|
|
|
|
|
|
2014-11-17 17:25:37 +00:00
|
|
|
def get_words(fname):
|
|
|
|
lines = open(fname)
|
2014-11-17 15:09:10 +00:00
|
|
|
words = []
|
2014-11-17 15:38:15 +00:00
|
|
|
for word in lines:
|
|
|
|
words.append(word.strip())
|
2014-11-17 15:09:10 +00:00
|
|
|
|
|
|
|
lines.close()
|
|
|
|
random.seed(634634)
|
|
|
|
random.shuffle(words)
|
2014-11-17 17:25:37 +00:00
|
|
|
words = words[:2**15]
|
|
|
|
assert len(words) == len(set(words))
|
2014-11-17 15:09:10 +00:00
|
|
|
return words
|
2014-11-17 17:25:37 +00:00
|
|
|
|
|
|
|
# These read like alien races from a sci-fi book
|
2014-11-21 08:32:23 +00:00
|
|
|
GOOGLE_WORDLIST = get_words("words/google-ngram-list")
|
2014-11-17 17:25:37 +00:00
|
|
|
# current best list for the three word hash
|
|
|
|
WORDNET_LEMMAS = get_words("words/wordnet-list")
|
2014-11-17 00:03:10 +00:00
|
|
|
|
|
|
|
# Human friendly word list, taken directly from humanhash project
|
2014-11-17 17:25:37 +00:00
|
|
|
# these are the best words but there are not enough of
|
|
|
|
# them so we only use them for the six word hash
|
2014-11-17 00:03:10 +00:00
|
|
|
HUMAN_WORDLIST = (
|
|
|
|
'ack', 'alabama', 'alanine', 'alaska', 'alpha', 'angel', 'apart', 'april',
|
|
|
|
'arizona', 'arkansas', 'artist', 'asparagus', 'aspen', 'august', 'autumn',
|
|
|
|
'avocado', 'bacon', 'bakerloo', 'batman', 'beer', 'berlin', 'beryllium',
|
|
|
|
'black', 'blossom', 'blue', 'bluebird', 'bravo', 'bulldog', 'burger',
|
|
|
|
'butter', 'california', 'carbon', 'cardinal', 'carolina', 'carpet', 'cat',
|
|
|
|
'ceiling', 'charlie', 'chicken', 'coffee', 'cola', 'cold', 'colorado',
|
|
|
|
'comet', 'connecticut', 'crazy', 'cup', 'dakota', 'december', 'delaware',
|
|
|
|
'delta', 'diet', 'don', 'double', 'early', 'earth', 'east', 'echo',
|
|
|
|
'edward', 'eight', 'eighteen', 'eleven', 'emma', 'enemy', 'equal',
|
|
|
|
'failed', 'fanta', 'fifteen', 'fillet', 'finch', 'fish', 'five', 'fix',
|
|
|
|
'floor', 'florida', 'football', 'four', 'fourteen', 'foxtrot', 'freddie',
|
|
|
|
'friend', 'fruit', 'gee', 'georgia', 'glucose', 'golf', 'green', 'grey',
|
|
|
|
'hamper', 'happy', 'harry', 'hawaii', 'helium', 'high', 'hot', 'hotel',
|
|
|
|
'hydrogen', 'idaho', 'illinois', 'india', 'indigo', 'ink', 'iowa',
|
|
|
|
'island', 'item', 'jersey', 'jig', 'johnny', 'juliet', 'july', 'jupiter',
|
|
|
|
'kansas', 'kentucky', 'kilo', 'king', 'kitten', 'lactose', 'lake', 'lamp',
|
|
|
|
'lemon', 'leopard', 'lima', 'lion', 'lithium', 'london', 'louisiana',
|
|
|
|
'low', 'magazine', 'magnesium', 'maine', 'mango', 'march', 'mars',
|
|
|
|
'maryland', 'massachusetts', 'may', 'mexico', 'michigan', 'mike',
|
|
|
|
'minnesota', 'mirror', 'mississippi', 'missouri', 'mobile', 'mockingbird',
|
|
|
|
'monkey', 'montana', 'moon', 'mountain', 'muppet', 'music', 'nebraska',
|
|
|
|
'neptune', 'network', 'nevada', 'nine', 'nineteen', 'nitrogen', 'north',
|
|
|
|
'november', 'nuts', 'october', 'ohio', 'oklahoma', 'one', 'orange',
|
|
|
|
'oranges', 'oregon', 'oscar', 'oven', 'oxygen', 'papa', 'paris', 'pasta',
|
|
|
|
'pennsylvania', 'pip', 'pizza', 'pluto', 'potato', 'princess', 'purple',
|
|
|
|
'quebec', 'queen', 'quiet', 'red', 'river', 'robert', 'robin', 'romeo',
|
|
|
|
'rugby', 'sad', 'salami', 'saturn', 'september', 'seven', 'seventeen',
|
|
|
|
'shade', 'sierra', 'single', 'sink', 'six', 'sixteen', 'skylark', 'snake',
|
|
|
|
'social', 'sodium', 'solar', 'south', 'spaghetti', 'speaker', 'spring',
|
|
|
|
'stairway', 'steak', 'stream', 'summer', 'sweet', 'table', 'tango', 'ten',
|
|
|
|
'tennessee', 'tennis', 'texas', 'thirteen', 'three', 'timing', 'triple',
|
|
|
|
'twelve', 'twenty', 'two', 'uncle', 'undress', 'uniform', 'uranus', 'utah',
|
|
|
|
'vegan', 'venus', 'vermont', 'victor', 'video', 'violet', 'virginia',
|
|
|
|
'washington', 'west', 'whiskey', 'white', 'william', 'winner', 'winter',
|
|
|
|
'wisconsin', 'wolfram', 'wyoming', 'xray', 'yankee', 'yellow', 'zebra',
|
|
|
|
'zulu')
|
|
|
|
|
|
|
|
|
|
|
|
class WordHasher(object):
|
|
|
|
def __init__(self):
|
|
|
|
"""Convert latitude and longitudes into human readable strings."""
|
|
|
|
self._symbols = "0123456789bcdefghjkmnpqrstuvwxyz"
|
|
|
|
self._decode_symbols = dict((ch, i) for (i, ch) in enumerate(self._symbols))
|
|
|
|
self._encode_symbols = dict((i, ch) for (i, ch) in enumerate(self._symbols))
|
2014-11-21 08:23:07 +00:00
|
|
|
self.six_wordlist = HUMAN_WORDLIST
|
|
|
|
self.three_wordlist = GOOGLE_WORDLIST
|
2014-11-17 00:03:10 +00:00
|
|
|
|
2024-06-28 13:35:43 +00:00
|
|
|
def three_words(self, lat_long):
|
2014-11-17 00:03:10 +00:00
|
|
|
"""Convert coordinate to a combination of three words
|
|
|
|
|
|
|
|
The coordinate is defined by latitude and longitude
|
|
|
|
in degrees.
|
|
|
|
"""
|
2024-06-28 13:35:43 +00:00
|
|
|
lat, lon = lat_long
|
2014-11-17 00:03:10 +00:00
|
|
|
gh = geohash.encode(lat, lon, 9)
|
2014-11-21 08:23:07 +00:00
|
|
|
words = "-".join(self.three_wordlist[p] for p in self.to_rugbits(self.geo_to_int(gh)))
|
2014-11-17 00:03:10 +00:00
|
|
|
return words
|
|
|
|
|
2024-06-28 13:35:43 +00:00
|
|
|
def six_words(self, lat_long):
|
2014-11-17 00:03:10 +00:00
|
|
|
"""Convert coordinate to a combination of six words
|
|
|
|
|
|
|
|
The coordinate is defined by latitude and longitude
|
|
|
|
in degrees.
|
|
|
|
|
|
|
|
With six words the word list contains only words
|
|
|
|
which are short, easy to pronounce and easy distinguish.
|
|
|
|
"""
|
2024-06-28 13:35:43 +00:00
|
|
|
lat, lon = lat_long
|
2014-11-17 00:03:10 +00:00
|
|
|
gh = geohash.encode(lat, lon, 9)
|
2014-11-21 08:23:07 +00:00
|
|
|
words = "-".join(self.six_wordlist[p] for p in self.to_bytes(self.pad(gh)))
|
2014-11-17 00:03:10 +00:00
|
|
|
return words
|
|
|
|
|
|
|
|
def decode(self, words):
|
|
|
|
"""Decode words back to latitude and longitude"""
|
|
|
|
words = words.split("-")
|
|
|
|
if len(words) == 3:
|
2014-11-21 08:23:07 +00:00
|
|
|
i = self.rugbits_to_int([self.three_wordlist.index(w) for w in words])
|
2014-11-17 00:03:10 +00:00
|
|
|
|
|
|
|
elif len(words) == 6:
|
2014-11-21 08:23:07 +00:00
|
|
|
i = self.bytes_to_int([self.six_wordlist.index(w) for w in words])
|
2014-11-17 00:03:10 +00:00
|
|
|
i = self.unpad(i)
|
|
|
|
|
|
|
|
else:
|
2014-11-17 00:10:43 +00:00
|
|
|
raise RuntimeError("Do not know how to decode a set of %i words."%(len(words)))
|
2014-11-17 00:03:10 +00:00
|
|
|
|
|
|
|
geo_hash = self.int_to_geo(i)
|
|
|
|
return geohash.decode(geo_hash)
|
|
|
|
|
|
|
|
def geo_to_int(self, geo_hash):
|
|
|
|
"""Decode `geo_hash` to an integer"""
|
|
|
|
base = len(self._symbols)
|
|
|
|
number = 0
|
|
|
|
for symbol in geo_hash:
|
|
|
|
number = number*base + self._decode_symbols[symbol]
|
|
|
|
|
|
|
|
return number
|
|
|
|
|
|
|
|
def int_to_geo(self, integer):
|
|
|
|
"""Encode `integer` to a geo hash"""
|
|
|
|
base = len(self._symbols)
|
|
|
|
symbols = []
|
|
|
|
while integer > 0:
|
|
|
|
remainder = integer % base
|
|
|
|
integer //= base
|
|
|
|
symbols.append(self._encode_symbols[remainder])
|
|
|
|
|
|
|
|
return ''.join(reversed(symbols))
|
|
|
|
|
|
|
|
def pad(self, geo_hash):
|
|
|
|
"""Pad nine character `geo_hash` to 48bit integer"""
|
|
|
|
assert len(geo_hash) == 9
|
|
|
|
return self.geo_to_int(geo_hash) * 8
|
|
|
|
|
|
|
|
def unpad(self, integer):
|
|
|
|
"""Remove 3bit of padding to get 45bit geo hash"""
|
|
|
|
return integer>>3
|
|
|
|
|
|
|
|
def to_bytes(self, integer):
|
|
|
|
"""Convert a 48bit `integer` to a list of 6bytes"""
|
|
|
|
bytes = [integer & 0b11111111]
|
|
|
|
for n in xrange(1,6):
|
|
|
|
div = 2**(n*8)
|
|
|
|
bytes.append((integer/div) & 0b11111111)
|
|
|
|
|
|
|
|
bytes.reverse()
|
|
|
|
return bytes
|
|
|
|
|
|
|
|
def bytes_to_int(self, bytes):
|
|
|
|
"""Convert a list of 6`bytes` to an integer"""
|
|
|
|
assert len(bytes) == 6
|
2014-11-17 09:45:51 +00:00
|
|
|
N = 0
|
|
|
|
bytes.reverse()
|
|
|
|
for n,b in enumerate(bytes):
|
|
|
|
N += b * (2**(8*(n)))
|
|
|
|
|
|
|
|
return N
|
2014-11-17 00:03:10 +00:00
|
|
|
|
|
|
|
def to_rugbits(self, integer):
|
|
|
|
"""Convert a 45bit `integer` to a list of 3rugbits
|
|
|
|
|
|
|
|
A rugbit is like a byte but with 15bits instead of eight.
|
|
|
|
"""
|
|
|
|
fifteen_bits = 0b111111111111111
|
|
|
|
rugbits = [(integer/(2**30)) & fifteen_bits,
|
|
|
|
(integer/(2**15)) & fifteen_bits,
|
|
|
|
integer & fifteen_bits]
|
|
|
|
return rugbits
|
|
|
|
|
|
|
|
def rugbits_to_int(self, rugbits):
|
|
|
|
"""Convert a list of `rugbits` to an integer"""
|
2014-11-17 09:45:51 +00:00
|
|
|
assert len(rugbits) == 3
|
2014-11-17 00:03:10 +00:00
|
|
|
return (rugbits[0] *(2**30)) + (rugbits[1] *(2**15)) + (rugbits[2])
|
|
|
|
|
|
|
|
|
|
|
|
DEFAULT_HASHER = WordHasher()
|
|
|
|
three_words = DEFAULT_HASHER.three_words
|
|
|
|
six_words = DEFAULT_HASHER.six_words
|
|
|
|
decode = DEFAULT_HASHER.decode
|