def calculate_Geohash(mbr: Rectangle) -> int:
try:
import pymorton as pm
except ImportError:
raise RuntimeError("Please install pymorton to calculate geohashes: pip install pymorton")
x, y = mbr.centroid()
return pm.interleave_latlng(y, x)
def test_pymorton(self):
# Test standard hashing
assert (pm.interleave(100, 50) == pm.interleave2(100, 50))
assert (pm.interleave(10, 50, 40) == pm.interleave3(10, 50, 40))
assert ((100, 30) == pm.deinterleave2(pm.interleave(100, 30)))
assert (pm.deinterleave2(pm.interleave(100, 30)) == (100, 30))
assert (pm.deinterleave3(pm.interleave(100, 30, 50)) == (100, 30, 50))
# Test geo-hashing
lat, lng = 40.712014, -74.008164
latlng_morton = pm.interleave_latlng(lat, lng)
assert('03023211232311330231120312032231' == \
latlng_morton)
assert(pm.deinterleave_latlng(latlng_morton) == \
(40.712013971060514, -74.00816400535405))
def test_pymorton(self):
# Test standard hashing
assert(pm.interleave(100, 50) == pm.interleave2(100, 50))
assert(pm.interleave(10, 50, 40) == pm.interleave3(10, 50, 40))
assert((100, 30) == pm.deinterleave2(pm.interleave(100, 30)))
assert(pm.deinterleave2(pm.interleave(100,30)) == (100, 30))
assert(pm.deinterleave3(pm.interleave(100, 30, 50)) == (100, 30, 50))
# Test geo-hashing
lat, lng = 40.712014, -74.008164
latlng_morton = pm.interleave_latlng(lat, lng)
assert('03023211232311330231120312032231' == \
latlng_morton)
assert(pm.deinterleave_latlng(latlng_morton) == \
(40.712013971060514, -74.00816400535405))
def convert_to_1D(location):
''' Map any store address to a single number! '''
if location.has_key('address'):
formatted_address, lat, lon = geo(location['address'])
lat = float(lat)
lon = float(lon)
elif location.has_key('lat') and location.has_key('lon'):
lat = float(location['lat'])
lon = float(location['lon'])
else:
return None
# Next, convert lat and long each into a positive int.
code = interleave_latlng(lat, lon)
# Now, turn the lat/long pair into a single number
if location.has_key('address'):
return {'id': code, 'lat': lat, 'lon': lon}
else:
return code
def test_geohash_ordinality(self):
assert pm.interleave_latlng(-40.723471, -73.985361) < pm.interleave_latlng(-40.523471, -73.785361)
def test_invalid_negative_latitude(self):
lat, lng = -220.712013, -74.008164
assert pm.deinterleave_latlng(pm.interleave_latlng(lat, lng)) == (round(lat + 180.0, 6), lng)
def test_invalid_negative_longitude(self):
lat, lng = -40.712013, -190.008164
assert pm.deinterleave_latlng(pm.interleave_latlng(lat, lng)) == (lat, round(lng + 180.0, 6))
def test_invalid_positive_longitude(self):
lat, lng = 40.712013, 190.008164
assert pm.deinterleave_latlng(pm.interleave_latlng(lat, lng)) == (lat, round(lng - 180.0, 6))
def test_standard_geohashing(self):
lat, lng = 40.712014, -74.008164
latlng_morton = pm.interleave_latlng(lat, lng)
assert '03023211232311330231120312032231' == latlng_morton
assert pm.deinterleave_latlng(latlng_morton) == (40.712014, -74.008164)
def test_ordering(self):
assert pm.interleave_latlng(-40.723471, -73.985361) < \
pm.interleave_latlng(-40.523471, -73.785361)
assert pm.interleave_latlng(40.723471, 40.734567) > \
pm.interleave_latlng(40.723271, 40.734367)
def Zorder_auto(filename, m, N):
ori = []
sample = []
with open(filename, 'r', encoding='utf8') as f:
for i in f:
ori = json.loads(i)
for index, i in enumerate(ori):
i['id'] = index
geo_hash_dict = {}
geo_hash = []
s_set = set()
datasets = {}
for index, i in enumerate(ori):
geohash = pm.interleave_latlng(i['lng'], i['lat'])
geo_hash_dict[str(index)] = geohash
datasets[str(index)] = [geohash]
geo_hash.append(geo_hash)
s_set.add(geohash)
# print(geo_hash)
# print(len(geo_hash))
# print(len(s_set))
a1 = sorted(geo_hash_dict.items(), key=lambda x: x[1], reverse=True)
aa = 20000
num = N
num1 = math.floor(aa / num)
if (aa - num * num1) % num1 == 0:
num += math.floor((aa - num * num1) / num1)
else:
num += math.floor((aa - num * num1) / num1) + 1
sample_list = []
z_lsit = []
attribute_dict = {}
attribute_list = []
x = np.zeros([num, 1])
for i in range(num):
if abs(aa - 1 - i * num1) < num1:
b = a1[i * num1:aa]
else:
b = a1[i * num1:(i + 1) * num1]
z_lsit.append(b)
# print(b, i)
c = random.sample(b, 1)
sample_list.append(c[0][0])
A = [{
"lng": ori[int(d)]['lng'],
"lat": ori[int(d)]['lat'],
"value": ori[int(d)]["value"],
'id': ori[int(d)]['id'],
'type': ori[int(d)]['type']
} for d in sample_list]
m.fit(A)
transform = [m.type_idx(i) for i in range(len(A))]
with open("./z-order/order_crime_output_{}.json".format(m.k),
mode='w',
encoding='utf8') as f:
res = []
for i in range(len(A)):
neighbors = m.neighbors[i]
res.append({
'id': A[i]['id'],
"type": transform[i],
"lat": A[i]["lat"],
"lng": A[i]["lng"],
"value": A[i]["value"],
"mx": m.score[i][0],
"my": m.score[i][1],
"neighbors": neighbors
})
json.dump(res, f)
with open('./z-order/order_crime_output_{}.json'.format(m.k),
'r',
encoding='utf-8') as f:
for i in f:
RecomputingPoints = json.loads(i)
oriData = A
correct = 0
mean = 0
for i in range(len(oriData)):
mean += RecomputingPoints[i]['value']
if RecomputingPoints[i]['type'] == ori[RecomputingPoints[i]
['id']]['type']:
correct += 1
return A, correct, RecomputingPoints, mean / (len(oriData))
def Zorder_auto_voting(filename, m, N, alpha, dirname, outputname):
ori = []
sample = []
with open(filename, 'r', encoding='utf8') as f:
ori = json.load(f)
for index, i in enumerate(ori):
i['id'] = index
geo_hash_dict = {}
geo_hash = []
s_set = set()
datasets = {}
for index, i in enumerate(ori):
geohash = pm.interleave_latlng(i['lng'], i['lat'])
geo_hash_dict[str(index)] = geohash
datasets[str(index)] = [geohash]
geo_hash.append(geo_hash)
s_set.add(geohash)
# print(geo_hash)
# print(len(geo_hash))
# print(len(s_set))
a1 = sorted(geo_hash_dict.items(), key=lambda x: x[1], reverse=True)
aa = len(ori)
num = math.floor(aa * float(N))
num1 = math.floor(aa / num)
if (aa - num * num1) % num1 == 0:
num += math.floor((aa - num * num1) / num1)
else:
num += math.floor((aa - num * num1) / num1) + 1
sample_list = []
z_lsit = []
x = np.zeros([num, 1])
for i in range(num):
if abs(aa - 1 - i * num1) < num1:
b = a1[i * num1:aa]
else:
b = a1[i * num1:(i + 1) * num1]
z_lsit.append(b)
# print(b, i)
c = random.sample(b, 1)
sample_list.append(c[0][0])
with open("../../../storage/zorder_temp.json", mode='w',
encoding='utf8') as f:
obj = []
for g in z_lsit:
obj.append([int(e[0]) for e in g])
json.dump(obj, f)
# z_list包含每个分组，统计每个分组的中心点，类比泊松盘，找到10个邻近分组。并获取每个分组内高低值的点分布
grouping_number = len(z_lsit)
center_list = []
c_list = []
diskCoverPoints = []
M_list = []
A_list = []
for i in z_lsit:
x = 0
y = 0
H_list = []
L_list = []
HH, HL, LL, LH = [], [], [], []
NH = []
NL = []
for j in i:
if ori[int(j[0])]['type'] == 'HH' or ori[int(
j[0])]['type'] == 'HL':
H_list.append(int(j[0]))
if ori[int(j[0])]['type'] == 'HH':
HH.append(int(j[0]))
else:
HL.append(int(j[0]))
elif ori[int(j[0])]['type'] == 'LL' or ori[int(
j[0])]['type'] == 'LH':
L_list.append(int(j[0]))
if ori[int(j[0])]['type'] == 'LL':
LL.append(int(j[0]))
else:
LH.append(int(j[0]))
if ori[int(j[0])]['type'] == 'HH' or ori[int(
j[0])]['type'] == 'LH':
NH.append(int(j[0]))
elif ori[int(j[0])]['type'] == 'LL' or ori[int(
j[0])]['type'] == 'HL':
NL.append(int(j[0]))
x += ori[int(j[0])]['lat']
y += ori[int(j[0])]['lng']
center_list.append([x / len(i), y / len(i)])
c_list.append([H_list, L_list])
M_list.append([HH, HL, LL, LH])
A_list.append([NH, NL])
# print(c_list)
# 找到邻接分组
nearbyM = {}
for i in range(grouping_number):
temp = {}
for j in range(grouping_number):
if i == j:
continue
else:
temp[str(j)] = math.sqrt(
(center_list[j][0] - center_list[i][0])**2 +
(center_list[j][1] - center_list[i][1])**2)
a1 = sorted(temp.items(), key=lambda x: x[1], reverse=False)
a1 = [int(d[0]) for d in a1[0:8]]
nearbyM[str(i)] = a1
# 找到相邻盘中包含自己的盘
bynearM = {}
# print(grouping_number)
for i in range(grouping_number):
temp = []
for j in nearbyM:
if i == int(j):
continue
else:
if i in nearbyM[j]:
temp.append(int(j))
bynearM[str(i)] = temp
# print(bynearM)
with open(dirname + '\\nearbyM.json', 'w', encoding='utf8') as f:
f.write(json.dumps(nearbyM))
with open(dirname + '\\bynearM.json', 'w', encoding='utf8') as f:
f.write(json.dumps(bynearM))
# nearbyM = []
# bynearM = []
# with open('./z-order/nearbyM.json', 'r', encoding='utf8') as f:
# for i in f:
# nearbyM = json.loads(i)
# with open('./z-order/bynearM.json', 'r', encoding='utf8') as f:
# for i in f:
# bynearM = json.loads(i)
with open(dirname + '\\nearbyM1.json', 'w', encoding='utf8') as f:
f.write(json.dumps(M_list))
# print(M_list)
a = update2(ori, c_list, nearbyM, bynearM, M_list, A_list, alpha)
A = [{
"lng": d['lng'],
"lat": d['lat'],
"value": d["value"],
'id': d['id'],
'type': d['type']
} for d in a]
m.fit(A)
transform = [m.type_idx(i) for i in range(len(A))]
with open(outputname, mode='w', encoding='utf8') as f:
res = []
for i in range(len(A)):
neighbors = m.neighbors[i]
res.append({
'id': A[i]['id'],
"type": transform[i],
"lat": A[i]["lat"],
"lng": A[i]["lng"],
"value": A[i]["value"],
"mx": m.score[i][0],
"my": m.score[i][1],
"neighbors": neighbors
})
json.dump(res, f)
with open(outputname, 'r', encoding='utf-8') as f:
for i in f:
RecomputingPoints = json.loads(i)
oriData = A
correct = 0
mean = 0
for i in range(len(oriData)):
mean += RecomputingPoints[i]['value']
if RecomputingPoints[i]['type'] == ori[RecomputingPoints[i]
['id']]['type']:
correct += 1
return A, correct, RecomputingPoints, ori, mean / len(oriData)
