-
Notifications
You must be signed in to change notification settings - Fork 0
/
names.py
521 lines (418 loc) · 19.3 KB
/
names.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
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
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
# python3 names.py dummy-sample.tab
# This script anonymizes the comment fields ('omschrijving') in Dutch bank transactions.
# It takes as input file a tab-separated text file with the following columns:
# MINISTERIE, BOEKJAAR, NAAM LEVERANCIER, OMSCHRIJVING, BEDRAG, VALUTA, GB_DATUM, EUR_BEDRAG
# It uses a number of external resources:
# - a list of 10,000 Dutch surnames and prefixes (‘de’, ‘ter’, ‘van’ etc.). Downloaded from http://www.naamkunde.net/?page_id=294
# - a list of 9,755 Dutch first names. Downloaded from http://www.naamkunde.net/?page_id=293
# - a list of 381,292 Dutch words. The file DFW.CD from the CELEX database https://catalog.ldc.upenn.edu/ldc96l14
# - a list of abbreviations, extracted from the transaction data itself: all words of 2–4 words that consist of only capital letters and are not a prefix or salutation (‘DHR’, ‘MEVR’, etc.), and occur at least 50 and that times in the data.5
# The generated output is a tab-separated file with the following 3 columns:
# item id of original row, anonymized omschrijving (names replaced by ***), list of found names
import sys
import re
#import operator
import json
import os
import xml.etree.ElementTree as ET
from random import randint
no_of_lines_read = 500000
datafilename = sys.argv[1]
# the json files are stored as index of the data. If a large dataset is analyzed multiple times, it is only read once and
# the following times the data is recovered from the json files
json_ngrams_filename = "ngrams.json"
json_inverted_index = "ngram_inverted_index.json"
json_data_columns = "data_columns.json"
records_with_names_filename = "anonymized_omschrijvingen+names.out.tab"
maxn = 5
def tokenize(t):
#text = t.lower()
# note that this function is very different from common tokenization functions because we need punctuation for recognizing names
text = t
text = re.sub("\n"," ",text)
text = re.sub(r'<[^>]+>',"",text) # remove all html markup
text = re.sub("\.",". ",text) # let . follow by whitespace for cases such as H.Ruissen
text = re.sub("-","- ",text) # let . follow by whitespace for cases such as Advies-A.W. van Engen
text = re.sub('[^a-zèéeêëėęûüùúūôöòóõœøîïíīįìàáâäæãåçćč&@#A-ZÇĆČÉÈÊËĒĘÛÜÙÚŪÔÖÒÓŒØŌÕÎÏÍĪĮÌ0-9/() \',.]', "", text)
wrds = text.split()
#print (wrds)
return wrds
months = ("jan","januari","feb","februari","mrt","maart","apr","april","mei","jun","juni","jul","juli","aug","augustus",
"sept","september","okt","oktober","nov","november","dec","december")
prefixes = set()
def initcaps(ngram):
ngram_initcaps = ""
words = ngram.split(" ")
for word in words:
#print(word)
if word.lower() not in prefixes:
ngram_initcaps += word.title() + " "
else:
#print(word.lower())
ngram_initcaps += word.lower() + " "
ngram_initcaps = ngram_initcaps.rstrip()
#print(ngram,ngram_initcaps)
return ngram_initcaps
print ("Read voornamenlijst")
firstnames = dict()
with open('voornamen_10kw.txt','r',encoding="UTF-8") as voornamenlijst:
for line in voornamenlijst:
line = re.sub("\r\n", "", line)
columns = line.rstrip().split("\t")
name = columns[0]
freq = columns[-1]
name = re.sub(" \([MV]\)","",name)
if name in firstnames:
firstnames[name] += int(freq)
else:
firstnames[name] = int(freq)
print("Read abbreviations list (capitalized words with length 2-4, frequency >= 50 in this data)")
abbreviations = dict()
with open('abbrev_freq.txt', 'r', encoding="UTF-8") as abbrevlist:
for line in abbrevlist:
line = re.sub("\r\n", "", line)
columns = line.rstrip().split("\t")
abbrev = columns[0]
freq = columns[-1]
if re.match("[0-9]",freq):
if int(freq) >= 50:
abbreviations[abbrev] = int(freq)
#for abbrev in abbreviations:
# print (abbrev,abbreviations[abbrev])
print ("Read achternamenlijst")
lastnames = dict()
tree = ET.parse('familienamen_10kw.xml')
root = tree.getroot()
for record in root.iter('record'):
if record is not None:
name = record.find('naam')
prefix = record.find('prefix')
freq = record.find('n2007')
if name.text is not None:
achternaam = name.text
if freq.text is not None:
lastnames[achternaam] = int(freq.text)
if prefix.text is not None:
#print (prefix.text, name.text)
achternaam = prefix.text+" "+name.text
lastnames[achternaam] = int(freq.text)
prefix_items = prefix.text.split(" ")
for item in prefix_items:
prefixes.add(item)
#print (achternaam,lastnames[achternaam])
prefixes.add("der")
prefixes.add("vd")
prefixes.add("v")
prefixes.add("d")
prefixes.add("'t")
#print(prefixes)
#for achternaam in lastnames:
# print (achternaam,lastnames[achternaam])
print ("Read CELEX frequencies")
celex_words = dict()
with open('DFW.CD','r',encoding="UTF-8") as celex:
for line in celex:
columns = line.rstrip().split('\\')
word = columns[1]
freq = int(columns[5])
#print(word,freq)
if word in celex_words:
if freq > celex_words[word]:
celex_words[word] = freq
else:
celex_words[word] = freq
def get_all_ngrams (text,maxn) :
words = tokenize(text)
i=0
terms = dict()
for word in words :
if len(word) > 1 and '@' not in word:
if word in terms :
terms[word] += 1
else :
terms[word] = 1
if len(words) >= 2 and maxn >= 2 :
if i< len(words)-1 :
#if words[i] not in stoplist and words[i+1] not in stoplist and words[i+1] != words[i]:
if words[i+1] != words[i]:
bigram = str(words[i])+ " " +str(words[i+1])
if bigram in terms :
terms[bigram] += 1
else :
terms[bigram] = 1
if len(words) >= 3 and maxn >= 3 :
if i < len(words)-2 :
if words[i+1] != words[i]:
trigram = str(words[i])+ " " +str(words[i+1])+ " " +str(words[i+2])
if trigram in terms :
terms[trigram] += 1
else :
terms[trigram] = 1
if len(words) >= 4 and maxn >= 4 :
if i < len(words)-3 :
if words[i+1] != words[i] and words[i+3] != words[i+2]:
fourgram = str(words[i])+ " " +str(words[i+1])+ " " +str(words[i+2])+ " " +str(words[i+3])
if fourgram in terms :
terms[fourgram] += 1
else :
terms[fourgram] = 1
if len(words) >= 4 and maxn >= 4 :
if i < len(words)-4 :
if words[i+1] != words[i] and words[i+3] != words[i+2] and words[i+4] != words[i+3]:
fivegram = str(words[i])+ " " +str(words[i+1])+ " " +str(words[i+2])+ " " +str(words[i+3])+ " " +str(words[i+4])
if fivegram in terms :
terms[fivegram] += 1
else :
terms[fivegram] = 1
i += 1
return terms
def filter_ngrams(freq_dict):
filtered_freq_dict = dict()
for ngram in freq_dict:
if re.match("^[a-zA-Z ,'.-]+$",ngram):
filtered_freq_dict[ngram] = freq_dict[ngram]
return filtered_freq_dict
def remove_overlapping_terms(terms):
remove = set()
for term1 in terms:
for term2 in terms:
if term1 != term2 and term1 in term2:
#print ("remove",term1,"because we have",term2)
remove.add(term1)
selection = []
for term in terms:
if term not in remove:
selection.append(term)
return selection
def merge_partly_overlapping_terms(terms,item):
columns_for_this_item = data_columns[str(item)] # json does not allow integers as keys, so it was stored with strings
omschrijving = columns_for_this_item[3]
terms_after_merging = []
terms_merged = set()
for term1 in terms:
for term2 in terms:
if term1 != term2 and omschrijving.index(term1) < omschrijving.index(term2):
#words1 = term1.split(" ")
longest_subterm = ""
words2 = term2.split(" ")
for i in range(0, len(words2) + 1):
for j in range(i + 1, len(words2) + 1):
subterm2 = " ".join(words2[i:j])
if subterm2 in term1:
if len(subterm2) > len(longest_subterm):
longest_subterm = subterm2
if len(longest_subterm) > 0:
#print(term2, "partly overlaps with", term1, "(overlap:", longest_subterm, ")")
term2_tail = re.sub(longest_subterm,"",term2)
merged = term1+term2_tail
#print("MERGED",merged)
terms_after_merging.append(merged)
terms_merged.add(term1)
terms_merged.add(term2)
#print("REMOVE:",term1)
#print("REMOVE:",term2)
for term in terms:
if term not in terms_merged:
terms_after_merging.append(term)
return terms_after_merging
def potential_lastname(word):
if word not in abbreviations \
and word.lower() not in months \
and not re.match(".*\.", word) \
and re.match("^[A-Z].*", word) \
and word.lower() not in celex_words \
and not "Pdirekt" in word \
and not re.match("[A-Za-z]+kosten", word):
return True
initial = "^[A-Z]\.?$"
aanhef = "^(dhr|mevr|dr|drs|mr|mw|ir|ing|mrs|sr)\.?$"
def count_name_features(ngram):
global initial
global aanhef
namefeats = []
namescore = 0.0
words = ngram.split(" ")
#print (">",ngram)
feats_per_index = dict()
for i in range(0,len(words)+1):
for j in range (i+1,len(words)+1):
#if len(words) > i+j-2:
subterm = " ".join(words[i:j])
#print (">",subterm)
if subterm.title() in firstnames:
namefeats.append("firstname")
# print(subterm,"firstname")
for k in range(i, j):
feats_per_index[k] = "firstname"
elif re.sub(",","",subterm).title() in lastnames or (" " in subterm and re.sub(",","",subterm) in lastnames):
# gaat goed: Jong, P. de; W.DE RIJKE
# fp: boek
namefeats.append("lastname")
#print(subterm, "lastname")
for k in range(i,j):
feats_per_index[k] = "lastname"
elif re.sub("\.$","",subterm.lower()) in prefixes:
namefeats.append("prefix")
#print(subterm,"prefix")
for k in range(i,j):
feats_per_index[k] = "prefix"
elif re.match(initial,subterm):
namefeats.append("initial")
#print(subterm,"initial")
for k in range(i,j):
feats_per_index[k] = "initial"
elif re.match(aanhef,subterm,re.IGNORECASE):
#namefeats.append("aanhef")
#print(subterm,"aanhef")
for k in range(i,j):
feats_per_index[k] = "aanhef"
elif re.match("^[A-Z][A-Z][A-Z]?$",subterm) and subterm not in abbreviations:
namefeats.append("initials")
#print(subterm,"initial")
for k in range(i,j):
feats_per_index[k] = "initials"
elif " " not in subterm and re.match("[A-Z][a-z].+",subterm)\
and subterm.upper() not in abbreviations \
and subterm.lower() not in months \
and not re.match(".*\.",subterm) and subterm.lower() not in celex_words \
and not "Pdirekt" in subterm \
and not re.match("[A-Za-z]+kosten", subterm):
namefeats.append("non-word")
for k in range(i,j):
feats_per_index[k] = "non-word"
if len(feats_per_index) == len(words)-1 and len(feats_per_index) == 1:
for k in feats_per_index:
if feats_per_index[k]== 'initial' or feats_per_index[k]== 'firstname':
if len(words) >= k+2:
if potential_lastname(words[k+1]):
feats_per_index[k+1] = 'lastname?'
break
elif k> 0 and k-1 not in feats_per_index:
if potential_lastname(words[k-1]):
feats_per_index[k-1] = 'lastname?'
break
#print (words,feats_per_index)
if len(feats_per_index) == len(words) and len(ngram) > 3:
prefix_and_aanhef_only = True
for k in feats_per_index:
if feats_per_index[k] != 'prefix' and feats_per_index[k] != 'initial' and feats_per_index[k] != 'aanhef':
prefix_and_aanhef_only = False
# print ("Prefix and aanhef only:",ngram)
if not prefix_and_aanhef_only:
#print ("NAME:",ngram)
namescore += len(namefeats)
return namefeats, namescore
'''
MAIN
'''
ngram_counts = dict()
inverted_index =dict()
data_columns=dict()
num_lines = 0
for line in open(datafilename,'r',encoding="UTF-8"):
num_lines += 1
if num_lines > 10002:
break
if num_lines > 10000 and os.path.isfile(json_ngrams_filename) and os.path.isfile(json_inverted_index) and os.path.isfile(json_data_columns):
print(json_ngrams_filename,"already exists: read from json")
with open(json_ngrams_filename,'r',encoding="UTF-8") as jsonfile:
ngram_counts = json.load(jsonfile)
print(json_inverted_index,"already exists: read from json")
with open(json_inverted_index,'r',encoding="UTF-8") as jsonfile:
inverted_index = json.load(jsonfile)
print(json_data_columns,"already exists: read from json")
with open(json_data_columns,'r',encoding="UTF-8") as jsonfile:
data_columns = json.load(jsonfile)
else:
print("Read", datafilename, ", read omschrijvingen into ngram dictionary, and save ngrams in inverted index")
i=0
with open(datafilename,'r',encoding='UTF-8') as data:
for line in data:
if i%1000 == 0:
sys.stderr.write(str(i)+" ")
sys.stderr.flush()
if i%10000 == 0:
sys.stderr.write('\n')
if i>=no_of_lines_read:
break
#MINISTERIE, BOEKJAAR, NAAM LEVERANCIER, OMSCHRIJVING, BEDRAG, VALUTA, GB_DATUM, EUR_BEDRAG,
line = re.sub("\r\n","",line)
#print(line)
if i > 0:
columns = line.rstrip().split("\t")
omschrijving = columns[3]
data_columns[str(i)] = columns # json does not allow integer keys, so store as string
ngrams = get_all_ngrams(omschrijving,maxn)
#print (i,omschrijving,ngrams)
ngrams = filter_ngrams(ngrams)
for ngram in ngrams:
if ngram in ngram_counts:
ngram_counts[ngram] += ngrams[ngram]
else:
ngram_counts[ngram] = ngrams[ngram]
items_for_this_ngram = []
if ngram in inverted_index:
items_for_this_ngram = inverted_index[ngram]
items_for_this_ngram.append(i)
inverted_index[ngram] = items_for_this_ngram
i +=1
with open (json_ngrams_filename,'w',encoding="UTF-8") as jsonfile:
json.dump(ngram_counts,jsonfile)
with open (json_inverted_index,'w',encoding="UTF-8") as jsonfile:
json.dump(inverted_index,jsonfile)
with open (json_data_columns,'w',encoding="UTF-8") as jsonfile:
json.dump(data_columns,jsonfile)
print("Count name evidence for",len(ngram_counts),"ngrams and store names per item")
names_per_item = dict()
all_names_uniq = dict() #key is name, value is evidence and score information
j = 0
for ngram in inverted_index:
if j % 10000 == 0:
sys.stderr.write(str(j) + " ")
sys.stderr.flush()
if j % 100000 == 0:
sys.stderr.write('\n')
j += 1
name_features, name_score = count_name_features(ngram)
jointevidence = "+".join(name_features)
#print(ngram,name_score)
if name_score >= 1.0:
items_for_this_ngram = inverted_index[ngram]
for item in items_for_this_ngram:
names_for_this_item = []
if item in names_per_item:
names_for_this_item = names_per_item[item]
names_for_this_item.append(ngram) # add this ngram to the array of names for this item
names_per_item[item] = names_for_this_item
for item in names_per_item:
#print ("#",item,": names for this item:",names_for_this_item)
names_without_overlap = remove_overlapping_terms(names_per_item[item])
names_merged = merge_partly_overlapping_terms(names_without_overlap,item)
names_per_item[item] = names_merged
print("Print items with names")
records_with_names_file = open(records_with_names_filename,'w',encoding="UTF-8")
records_with_names_file.write("id\tomschrijving anoniem\tnames (automatic)\tnames (manual)\n")
item_count = 0
count_no_names = 0
name_count = 0
for item in data_columns:
#print ("item:",item)
item_count +=1
columns_for_this_item = data_columns[str(item)] # json does not allow integers as keys, so it was stored with strings
omschrijving = columns_for_this_item[3]
omschrijving_anonymized = omschrijving
has_name = False
if int(item) in names_per_item:
for name in names_per_item[int(item)]:
#print (omschrijving,name)
omschrijving_anonymized = re.sub(name,"***",omschrijving_anonymized)
has_name = True
name_count += 1
records_with_names_file.write(str(item) + "\t" + omschrijving_anonymized + "\t" + str(names_per_item[int(item)])+"\n")
if not has_name:
records_with_names_file.write(str(item) + "\t" + omschrijving_anonymized+"\n")
count_no_names +=1
records_with_names_file.close()
print("total number of rows in the data:",item_count-1)
print("number of rows with at least one name:",item_count-count_no_names)
print("number of names found:",name_count)