/
process.py
323 lines (192 loc) · 11.9 KB
/
process.py
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
import sys
from time import perf_counter
sys.setrecursionlimit(10000)
from fuzzywuzzy import process,fuzz
from sklearn.feature_extraction.text import CountVectorizer
import pandas as pd
import pickle
import numpy as np
import requests
import os
import math
from sklearn.neighbors import BallTree
pd.set_option("display.max_columns",101)
def download_file_or_not(url=None,local_file=None,overwrite=False):
if not os.path.exists(local_file) or overwrite:
"""do download if local file doesn't exist or overwrite=True"""
request=requests.get(url=url)
with open(local_file,'w') as f:
f.write(request.text)
df=pd.read_csv(local_file)
return df
def concatenate_address_fields_or_not(df):
if not 'full_address' in df.columns:
df['full_address'] = df.Address.apply(lambda x: x.strip()) + ' ' + df['Unit Number'].fillna('').apply(
lambda x: str(x).strip()) + ' ' + df.Zipcode.apply(lambda x: str(x).strip())
df['full_address'] = df['full_address'].apply(lambda x: x.replace(' ', ' ').strip())
return df
def join_data(reference_path=None, target_path=None,save_path=None,radius=None,sample=0,head=False):
reference=pd.read_csv(reference_path)
target=pd.read_csv(target_path,low_memory=False)
reference['LonRad'] = reference['Longitude'].apply(math.radians)
reference['LatRad'] = reference['Latitude'].apply(math.radians)
if 'Permit Address' in target.columns:
target.rename(columns={'Permit Address':'Address'},inplace=True)
if head:
print(target.head(20))
#target.dropna(subset=['Location'],inplace=True)
#target.reset_index(drop=True,inplace=True)
if 'Longitude' not in target.columns:
target['Longitude']=target.Location.apply(lambda x: float(str(x[1:-1]).split(',')[1]) if not pd.isnull(x) else None)
target['Latitude']=target.Location.apply(lambda x: float(str(x[1:-1]).split(',')[0]) if not pd.isnull(x) else None)
target['LonRad']=target['Longitude'].apply(math.radians)
target['LatRad']=target['Latitude'].apply(math.radians)
start=perf_counter()
k=BallTree(data=reference[['LonRad','LatRad']],metric='haversine')
i=target#[target.index==108] #274171]
if sample>0:
i=i.sample(sample,replace=False)
#[['Longitude','Latitude']].iloc[5]
r_radians=radius/40075000*2*math.pi*.7
# print(rows[['LonRad','LatRat']].values.reshape(1,-1).isfinite())
# i['knn_indices_found']=i.apply(lambda rows:list(k.query_radius(rows[['LonRad','LatRad']].values.reshape(1,-1),r=r_radians)),axis=1)
# i.assign(knn_indices_found=k.query_radius(i[['LonRad','LatRad']].values.reshape(1,-1),r=r_radians))
# i.insert(len(i.columns),'knn_indices_found',k.query_radius(i[['LonRad','LatRad']].values,r=r_radians))
i['knn_indices_found']=i.apply(lambda cols:list(k.query_radius(cols[['LonRad','LatRad']].values.reshape(1,2),r=r_radians)) if not pd.isnull(cols['LonRad']) else None,axis=1).apply(lambda x: x[0] if x is not None else None)
i['found_count']=i.knn_indices_found.apply(lambda x:len(x) if x is not None else None).values
i['longitudes_found']=i.apply(lambda rows:list(reference['Longitude'].iloc[rows['knn_indices_found']]) if rows['knn_indices_found'] is not None else None,axis=1)
i['latitudes_found']=i.apply(lambda rows:list(reference['Latitude'].iloc[rows['knn_indices_found']]) if rows['knn_indices_found'] is not None else None ,axis=1)
i['addresses_found']=i.apply(lambda rows:list(reference['full_address'].iloc[rows['knn_indices_found']] ) if rows['knn_indices_found'] is not None else None,axis=1)
# i.assign(most_likely=process.extractOne(i['Address'],choices=i['addresses_found']) if not i.Address.isnull().item() else None)
#
i['most_likely']=i.apply(lambda cols:process.extractOne(cols['Address'],choices=cols['addresses_found']
if (not pd.isnull(cols['Address']) and cols['addresses_found'] is not None) else None),axis=1)
i['most_likely_address']=i.apply(lambda cols: cols['most_likely'][0] if cols['most_likely'] is not None else None,axis=1)
i['most_likely_index']=i.apply(lambda cols: cols['addresses_found'].index(cols['most_likely_address']) if (not pd.isnull(cols['Address']) and cols['most_likely_address'] is not None) else None,axis=1)
i['most_likely_lon']=i.apply(lambda cols:cols['longitudes_found'][int(cols['most_likely_index'])] if not pd.isnull(cols['most_likely_index']) else None,axis=1)
i['most_likely_lat'] = i.apply(lambda cols: cols['latitudes_found'][int(cols['most_likely_index'])] if not pd.isnull(cols['most_likely_index']) else None,axis=1)
i['most_likely_edit_ratio']=i['most_likely'].apply(lambda x: x[1] if x is not None else None)
i['master_indices']=i.apply(lambda rows: int(rows['knn_indices_found'][rows['addresses_found'].index(rows['most_likely_address'])]) if rows['most_likely'] is not None else None,axis=1)
i.to_csv(save_path)
return (i,radius)
def plot_one(s):
import math
import matplotlib.pyplot as plt
import matplotlib.ticker as ticker
i=s[0].sample(1)
radius=s[1]
print(i.columns)
import seaborn as sns
f,ax=plt.subplots()
#ax.annotate(i['Address'].values[0],xy=(i['Longitude'],i['Latitude']),xytext=(i['Longitude'],i['Latitude']))
up=1.000001
down=.999999
x=i['longitudes_found'].values[0]
y=i['latitudes_found'].values[0]
ax.scatter(x=x,y=y,marker='x')
#ax.scatter(x=i['Longitude'],y=i['Latitude'],label=None) # ,label='Inspection Address:{}'.format(i['Address'].values[0]))
circle=plt.Circle(xy=(i['Longitude'],i['Latitude']),radius=radius,color='r',fill=False)
ax.add_artist(circle)
ax.set_xlim(i['Longitude'].values+circle.radius,i['Longitude'].values-circle.radius)
ax.set_ylim(i['Latitude'].values+circle.radius,i['Latitude'].values-circle.radius)
best_index=int(i.most_likely_index)
# print(i.addresses_found.iloc(best_index)))
# [ax.annotate(address,xy=(x,y),xytext=(x,y),rotation=45) for address,x,y in zip(i.addresses_found.values[0],x,y)]
ax.annotate('Finding: {}'.format(i['Address'].values[0]),xy=(x[best_index],y[best_index]),xycoords='data',
xytext=(.5,1.05),textcoords='axes fraction',
arrowprops=dict(facecolor='black', shrink=0.05,width=1,connectionstyle='arc3,rad=.2'))
ax.annotate('Found: {}'.format(i['most_likely_address'].values[0]),xy=(x[best_index],y[best_index]),xycoords='data',
xytext=(x[best_index],y[best_index]),textcoords='data',
)
ax.yaxis.set_major_formatter(ticker.FormatStrFormatter('%.5f'))
ax.xaxis.set_major_formatter(ticker.FormatStrFormatter('%.5f'))
plt.legend()
plt.show()
def plot_goog(s,sample=1):
import webbrowser
"""['Inspection Number', 'Inspection Type', 'Inspection Type Description',
'Address', 'Inspection Address Zipcode', 'Battalion', 'Station Area',
'Fire Prevention District', 'Billable Inspection',
'Inspection Start Date', 'Inspection End Date', 'Inspection Status',
'Return Date', 'Corrective Action Date', 'Referral Agency',
'Complaint Number', 'Permit Number', 'Referral Number',
'Violation Number', 'DBI Application Number', 'Invoice Date',
'Second Notice Date', 'Final Notice Date', 'Lien Date',
'Sent to Bureau of Delinquent Revenue', 'Invoice Amount', 'Fee',
'Penalty Amount', 'Posting Fee', 'Interest Amount', 'Paid Amount',
'Paid Date', 'Supervisor District', 'Neighborhood District',
'Location', 'Longitude', 'Latitude', 'knn_indices_found', 'found_count',
'longitudes_found', 'latitudes_found', 'coordinates_found',
'addresses_found', 'most_likely', 'most_likely_address',
'most_likely_index', 'most_likely_edit_ratio', 'master_indices']"""
from string import Template
import math
inspections=s[0]
radius = s[1]
with open('map_template.html','r') as f:
template=Template(f.read())
for plots in range(sample):
i = inspections.sample(n=1,replace=False)
o=.0003
search_coord=('{{ lat:{},lng:{} }}'.format(i.Latitude.values[0],i.Longitude.values[0]))
offset=('{{ lat:{},lng:{} }}'.format(i.Latitude.values[0]+o,i.Longitude.values[0]+o))
target_address='"'+i['Address'].values[0]+'"'
mle_address='"'+i['most_likely_address'].values[0]+'"'
mle_lon=i['most_likely_lon'].values[0]
mle_lat=i['most_likely_lat'].values[0]
mle_coord = ('{{ lat:{},lng:{} }}'.format(mle_lat, mle_lon))
x = i['longitudes_found'].values[0]
y = i['latitudes_found'].values[0]
lontext=['lng:']*len(x)
lattext=['lat:']*len(y)
coordinates='['+','.join(['{{{}{},{}{}}}'.format(lontext,x,lattext,y) for lontext,x,lattext,y in list(zip(lontext,x,lattext,y))])+']'
labels = i['addresses_found'].values[0]
l=i['addresses_found'].values
title='Searching for {} \n -> Found {}'.format(target_address,mle_address).replace('"','')
text=template.substitute(radius=radius,coordinates=coordinates,labels=labels,target_address=target_address,search_coord=search_coord,offset=offset,mle_address=mle_address,mle_coord=mle_coord,title=title)
with open('map.html','w') as f:
f.write(text)
print(text)
#webbrowser.open_new('map.html')
#plot_goog(join_data(radius=60))
def analyze(path=None):
i=pd.read_csv(path)
results=i[['Address','most_likely_address','most_likely_edit_ratio','master_indices','found_count','addresses_found']]
print(results[results.isnull().any(axis=1)].count())
end=perf_counter()
#print('elapsed {} sec'.format((end-start)))
print(results[['Address','most_likely_address','most_likely_edit_ratio']][(results.most_likely_edit_ratio<60) & (results.most_likely_edit_ratio<=100)].sort_values('most_likely_edit_ratio',ascending=False))
import matplotlib.pyplot as plt
import seaborn as sns
f=plt.figure()
sns.distplot(results['most_likely_edit_ratio'][results.notnull().all(axis=1)],kde=False,bins=20)
plt.title='Histogram of edit ratios for {}'.format(target_file)
plt.savefig('Histogram of edit ratios for {}.png'.format(target_file))
# i.to_csv(os.path.join(processed_data_directory,'Inspections_with_Master_Address.csv'))
if __name__ == '__main__':
target_directory='../data/raw'
processed_data_directory='../data/processed'
reference='Addresses_-_Enterprise_Addressing_System_cleaned.csv'
reference_path=os.path.join(processed_data_directory,reference )
target_file='Police_Department_Incidents.csv'
target_path=os.path.join(target_directory,target_file)
save_path=os.path.join(processed_data_directory,'{}_with_matched.csv'.format(target_file.replace('.csv','')))
def modify_business():
df = pd.read_csv(os.path.join(data_directory, 'Registered_Business_Locations_-_San_Francisco.csv'))
stuff=df['Business Location'].astype(str).apply(lambda x: x.split('\n'))
df['Address']=stuff.apply(lambda x:x[0] if x is not None else None)
df['Location']=stuff.apply(lambda x:x[2] if len(x)>2 else None)
print(df.head(50))
exit()
print (df.Location.apply(lambda x:x is None).sum())
exit()
df = df.to_csv(os.path.join(data_directory, 'Registered_Business_Locations_-_San_Francisco.csv'))
def modify_labeled():
df=pd.read_csv(os.path.join(data_directory,'labeled-property-tax-addresses_060117.csv'))
df['Zipcode of Parcel']=df.apply(lambda cols: int(cols['Zipcode of Parcel']) if not pd.isnull(cols['Zipcode of Parcel']) else '',axis=1)
df['Address']=df.apply(lambda cols: '{} {} {} {}'.format(cols['Start Street Number'],cols['Street Name'],cols['Street Suffix'],cols['Zipcode of Parcel'] ),axis=1)
df.to_csv(os.path.join(data_directory,'labeled-property-tax-addresses_060117.csv'))
data=join_data(reference_path=reference_path,target_path=target_path,save_path=save_path,radius=40,sample=0,head=True)
analyze(save_path)
# plot_goog(data,2)
exit()