question3_compare_month.py 8.94 KB
Newer Older
Tianyang's avatar
Tianyang committed
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
import numpy as np
import matplotlib.pyplot as plt
from functools import reduce
from database_pre3 import connection
import matplotlib.pyplot as plt
import re
import folium
import random
import datetime


table_variable = [
    'date',
    'lon',
    'lat',
    'station',
    'time',
    'alti',
    'drct',
    'dwpf', 
    'feel',
    'gust',
    'ice_accretion_1hr',
    'ice_accretion_3hr',
    'ice_accretion_6hr',
    'metar',
    'mslp',
    'p01i',
    'peak_wind_drct',
    'peak_wind_gust',
    'peak_wind_time',
    'relh',
    'sknt',
    'skyc1',
    'skyc2',
    'skyc3',
    'skyc4',
    'skyl1',
    'skyl2',
    'skyl3',
    'skyl4',
    'tmpf',
    'vsby',
    'wxcodes']




def add (x,y):
    return x+y


def abs_diff(x,y):
    longth = len(x)
    result = []
    for i in range(longth):
        result.append(abs(x[i]-y[i]))
    total = 0
    for each in result:
        total += each
    return total

def diff(x,y):
    return x-y



#caculate mean reduce
#input [count,mean]
def reduceFonction (x,y):
    result = []
    for i in range(2):
        result.append(reduce(add,[x[i],y[i]]))
    return result

#input [valeur] -> [count,mean]
def mapFonction1 (x):
    return [1,x]

#input [count,mean] -> [mean]
def mapFonction2 (x):
    return x[1]/x[0]


#Map reduce to caculate the means of each month of each station
def mapReduce_kmeans(targetNB,start,end):
    results = dict()
    for row in session.execute("select * from caitiany.database_kmeans where date >= '%s' and date <= '%s' ALLOW FILTERING"%(start,end)):
        data_target = row[targetNB]
        if data_target == None:
            continue
        #We want to take out the month correspondant and add month into keys
        dateparser = re.compile("(\d+)-(?P<month>\d+)-(\d+) (?P<time>\d+:\d+)")
        match_month = dateparser.match(str(row[0]))
        time = match_month.groupdict()
        month = time["month"]
        data_espace = (row[1],row[2],row[3],month)
        if results.get(data_espace) is None:
            results[data_espace] = mapFonction1(data_target)
        else:
            mapresult = mapFonction1(data_target)
            results[data_espace] = reduceFonction(mapresult,results[data_espace])
    for eachEspace in results:
        results[eachEspace] = mapFonction2(results[eachEspace])
    #Now we need to put the data of each station all together
    newResult = dict()
    for each in results:
        newKey = each[0:3]
        if newResult.get(newKey) is None:
            newResult[newKey] = [results[each]]
        else:
            newResult[newKey] = newResult[newKey]+[results[each]]
    max = 0
    for each in newResult:
        if len(newResult[each]) >max:
            max = len(newResult[each])
    deletList = []
    #We want to check if there is a missing data so we check the number of month max
    #Then we delete all the station who have the missing data
    for each in newResult:
        if len(newResult[each]) < max:
            deletList.append(each)
    for delete in deletList:
        newResult.pop(delete)
    return newResult

#Caculate the difference of the number of the clusters 
def cluster_nb_diff(centre_new,centre):
    sum = 0
    for i in range(3):
        sum += abs(centre_new[i][0]-centre[i][0])
    return sum/3



#input [tmpt] -> [tmpt,tmpt,tmpt,mean]
def map1_kmeans(x):
    return [x,x,x,x]


#input [tmpt,tmpt,tmpt,tmpt] and [c1,c2,c3,0] -> [|tmpt - c1|,|tmpt - c2|,|tmpt - c3|,tmpt]
def reduceKmeans (x,y):
    result = []
    for i in range(3):
        result.append(reduce(abs_diff,[x[i],y[i]]))
    result.append(x[3])
    return result


#input [|tmpt - c1|,|tmpt - c2|,|tmpt - c3|,tmpt] -> [cluster number, min(|tmpt - c|), tmpt]
def map2_kmeans(x):
    min_value = 10000000000000
    index = 0
    for each in range(3):
        if min_value > x[each]:
            min_value = x[each]
            index = each
    return [index,min_value,x[3]]



#Update the new center by means
def MapnewCentre(x):
    result = []
    longth = len(x[1])
    for i in range(longth):
        if x[0] != 0:
            result.append(x[1][i]/x[0])
        else:
            result.append(0)
    return result

#caculate moyen
def caculateMoyen(x):
    total = 0
    count = 0
    for i in x:
        count += 1
        total += i
    return total/count



#caculate sum of two list
def caculateSumofList(x,y):
    if x == []:
        return y
    longth = len(x)
    result = []
    for i in range(longth):
        result.append(x[i] + y[i])
    return result


#The main algorithm of Kmeans
def kmeans (targetNB,target,start,end):
    #cluster est pour stocler lat, lon de chaque point de chaque cluster
    cluster = [[],[],[]]

    result = mapReduce_kmeans(targetNB,start,end)

    if len(result) < 3:
        raise Exception ("We\'ve just searched less than 3 station!!")

    #mettre ramdom 3 point comme le centres init
    center1 = random.choice(list(result))
    center2 = random.choice(list(result))
    center3 = random.choice(list(result))
    init_point_values = [result[center1],result[center2],result[center3]]
    #init_point_values = [result[i] for i in result.keys()][:3]
    init_point_keys = [center1,center2,center3]

    #3centre with [point count, temprature centre]
    centre = {0:[0,0],1:[0,0],2:[0,0]}
    centre_new = {0:[0,0],1:[0,0],2:[0,0]}

    for key in centre.keys():
        centre[key] = [1,init_point_values[key]]
        centre_new[key] = [1,init_point_values[key]]
        cluster[key].append(init_point_keys[key])
    

    #init the centre new and result new for mapreduce
    result_new = dict()
    #When the number of point of cluster don't change,stop
    while True:
        for eachkey in result:
            if eachkey in cluster[0] or eachkey in cluster[1] or eachkey in cluster[2]:
                continue
            
            #caculate the distance between the data of this lingne and the centre

            #Map1_kemeans
            result_new[eachkey] = map1_kmeans(result[eachkey])
            centre_values = [centre[0][1],centre[1][1],centre[2][1]]

            #Reduce
            result_new[eachkey] = reduceKmeans(result_new[eachkey],centre_values)

            #Map2_kmeans
            result_new[eachkey] = map2_kmeans(result_new[eachkey])

        #Put all the distance and points into the clusters
        #Result format [cluster number, min(|tmpt - c|),tmpt - c]
        for eachpoint in result_new:
            clusterNB = result_new[eachpoint][0]
            centre_new[clusterNB][0] += 1
            centre_new[clusterNB][1] = caculateSumofList(centre_new[clusterNB][1],result_new[eachpoint][2])
            cluster[clusterNB].append(eachpoint)
        #compare centre_new and centre, if
        if not cluster_nb_diff(centre_new,centre) > 1:
            break
        else:
            #caculate the new centre
            print ("Cluster number differ:  ",cluster_nb_diff(centre_new,centre))
            for eachculster in centre_new:
                centre_new[eachculster][1] = MapnewCentre(centre_new[eachculster])
            print("center new: ",centre_new)
            print("center old: ",centre)
            centre = centre_new
            centre_new = {0:[0,[]],1:[0,[]],2:[0,[]]}
            result_new = dict()
            cluster = [[],[],[]]
    createMap(cluster,result)



#Create the map of the cluster
def createMap (cluster,result):
    mean_lat = 0
    count = 0
    for each in [cluster[0],cluster[1],cluster[2]]:
        for each_pos in each:
            mean_lat += each_pos[0]
            count += 1
    mean_lat = mean_lat/count

    mean_lon = 0
    count = 0
    for each in [cluster[0],cluster[1],cluster[2]]:
        for each_pos in each:
            mean_lon += each_pos[1]
            count += 1
    mean_lon = mean_lon/count

    m = folium.Map(location=[mean_lon,mean_lat],zoom_start=6)

    color = {0:'blue',1:'red',2:'green'}
    i = 0
    for each in [cluster[0],cluster[1],cluster[2]]:
        for each_pos in each:
            label = str(each_pos[2])+"\n"+""+target+":"+str(round(caculateMoyen(result[(each_pos[0],each_pos[1],each_pos[2])]),2))
            folium.Marker([each_pos[1],each_pos[0]],
                        popup=label,
                        icon=folium.Icon(color=color[i])).add_to(m)
        i +=1
    m.save("Projet-NF26/map.html")
    print("Generate successfully")


#Check which number of the indicateur
def checkNBvariable (x):
    i=0
    for each in table_variable:
        if x == each:
            return i
        i += 1
    print ('Doesn\'t exist!!')


if __name__ == "__main__":
    session = connection()
    start = input("Please enter the start time [form: AAAA-MM-DD (From 2008-1-1 to 2017-12-30)]:  ")
    end = input("Please enter the end time [form: AAAA-MM-DD (From 2008-1-1 to 2017-12-30)]:  ")
    target = input("Which indicator do you want to check [tmpf,dwpf,etc]:  ")
    targetNB = checkNBvariable(target)
    #start = '2008-01-01'
    #end = '2013-12-12'
    #data = session.execute_async("select * from caitiany.database_kmeans where date >= '%s' and date <= '%s' ALLOW FILTERING"%(start,end))
    kmeans(targetNB,target,start,end)