This repository has been archived by the owner on Oct 3, 2019. It is now read-only.
-
Notifications
You must be signed in to change notification settings - Fork 8
/
matching.ex
692 lines (620 loc) · 20.6 KB
/
matching.ex
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
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
defmodule Zxcvbn.Matching do
@moduledoc false
alias Zxcvbn.Scoring
@l33t_table %{
a: ["4", "@"],
b: ["8"],
c: ["(", "{", "[", "<"],
e: ["3"],
g: ["6", "9"],
i: ["1", "!", "|"],
l: ["1", "|", "7"],
o: ["0"],
s: ["$", "5"],
t: ["+", "7"],
x: ["%"],
z: ["2"]
}
@regexen %{recent_year: ~r/19\d\d|200\d|201\d/}
@date_max_year 2050
@date_min_year 1000
@date_splits %{
# for length-4 strings, eg 1191 or 9111, two ways to split:
4 => [
# 1 1 91 (2nd split starts at index 1, 3rd at index 2)
[1, 2],
# 91 1 1
[2, 3]
],
5 => [
# 1 11 91
[1, 3],
# 11 1 91
[2, 3]
],
6 => [
# 1 1 1991
[1, 2],
# 11 11 91
[2, 4],
# 1991 1 1
[4, 5]
],
7 => [
# 1 11 1991
[1, 3],
# 11 1 1991
[2, 3],
# 1991 1 11
[4, 5],
# 1991 11 1
[4, 6]
],
8 => [
# 11 11 1991
[2, 4],
# 1991 11 11
[4, 6]
]
}
@shifted_rx ~r/[~!@#$%^&*()_+QWERTYUIOP{}|ASDFGHJKL:"ZXCVBNM<>?]/
@max_delta 5
# ----------------------------------------------------------------------------
# Matchers
# ----------------------------------------------------------------------------
@doc """
Omnimatch -- combine everything.
"""
def omnimatch(_password) do
# matches = []
# matchers = [
# @dictionary_match
# @reverse_dictionary_match
# @l33t_match
# @spatial_match
# @repeat_match
# @sequence_match
# @regex_match
# @date_match
# ]
# for matcher in matchers
# @extend matches, matcher.call(this, password)
# @sorted matches
end
@doc """
Dictionary match (common passwords, english, last names, etc).
"""
def dictionary_match(password, ranked_dictionaries \\ ranked_dictionaries())
def dictionary_match("", _), do: []
def dictionary_match(password, ranked_dictionaries) do
password
|> word_permutations
|> Enum.flat_map(&dictionaries_matches(ranked_dictionaries, &1, []))
|> sorted
end
defp word_permutations(word) do
len = String.length(word)
for i <- 0..len,
j <- 1..(len - i),
do: %{i: i, j: i + j - 1, token: String.slice(word, i, j)}
end
defp dictionaries_matches([], _, matches), do: matches
defp dictionaries_matches([head | tail], permutation, matches) do
matched_word = String.downcase(permutation.token)
case head.words[matched_word] do
nil ->
dictionaries_matches(tail, permutation, matches)
rank ->
match = %{
i: permutation.i,
j: permutation.j,
token: permutation.token,
matched_word: matched_word,
reversed: false,
l33t: false,
rank: rank,
dictionary_name: head.name,
pattern: "dictionary"
}
dictionaries_matches(tail, permutation, matches ++ [match])
end
end
def reverse_dictionary_match(password, ranked_dictionaries \\ ranked_dictionaries()) do
password
|> String.reverse()
|> dictionary_match(ranked_dictionaries)
|> Enum.map(&reverse_match(&1, String.length(password)))
|> sorted
end
defp reverse_match(match, length) do
%{
match
| i: length - 1 - match.j,
j: length - 1 - match.i,
reversed: true,
token: String.reverse(match.token)
}
end
@doc """
Dictionary match with common l33t substitutions.
"""
def l33t_match(
_password,
_ranked_dictionaries \\ ranked_dictionaries(),
_l33t_table \\ @l33t_table
) do
# matches = []
# for sub in @enumerate_l33t_subs @relevant_l33t_subtable(password, _l33t_table)
# break if @empty sub # corner case: password has no relevant subs.
# subbed_password = @translate password, sub
# for match in @dictionary_match(subbed_password, _ranked_dictionaries)
# token = password[match.i..match.j]
# if token.toLowerCase() == match.matched_word
# continue # only return the matches that contain an actual substitution
# match_sub = {} # subset of mappings in sub that are in use for this match
# for subbed_chr, chr of sub when token.indexOf(subbed_chr) != -1
# match_sub[subbed_chr] = chr
# match.l33t = true
# match.token = token
# match.sub = match_sub
# match.sub_display = ("#{k} -> #{v}" for k,v of match_sub).join(', ')
# matches.push match
# @sorted matches.filter (match) ->
# # filter single-character l33t matches to reduce noise.
# # otherwise '1' matches 'i', '4' matches 'a', both very common English words
# # with low dictionary rank.
# match.token.length > 1
end
@doc """
Spatial match (qwerty/dvorak/keypad).
"""
def spatial_match(_password, _graph \\ graphs()) do
# matches = []
# for graph_name, graph of _graphs
# @extend matches, @spatial_match_helper(password, graph, graph_name)
# @sorted matches
end
@doc """
Repeats (aaa, abcabcabc) and sequences (abcdef).
"""
def repeat_match() do
# matches = []
# greedy = /(.+)\1+/g
# lazy = /(.+?)\1+/g
# lazy_anchored = /^(.+?)\1+$/
# lastIndex = 0
# while lastIndex < password.length
# greedy.lastIndex = lazy.lastIndex = lastIndex
# greedy_match = greedy.exec password
# lazy_match = lazy.exec password
# break unless greedy_match?
# if greedy_match[0].length > lazy_match[0].length
# # greedy beats lazy for 'aabaab'
# # greedy: [aabaab, aab]
# # lazy: [aa, a]
# match = greedy_match
# # greedy's repeated string might itself be repeated, eg.
# # aabaab in aabaabaabaab.
# # run an anchored lazy match on greedy's repeated string
# # to find the shortest repeated string
# base_token = lazy_anchored.exec(match[0])[1]
# else
# # lazy beats greedy for 'aaaaa'
# # greedy: [aaaa, aa]
# # lazy: [aaaaa, a]
# match = lazy_match
# base_token = match[1]
# [i, j] = [match.index, match.index + match[0].length - 1]
# # recursively match and score the base string
# base_analysis = scoring.most_guessable_match_sequence(
# base_token
# @omnimatch base_token
# )
# base_matches = base_analysis.sequence
# base_guesses = base_analysis.guesses
# matches.push
# pattern: 'repeat'
# i: i
# j: j
# token: match[0]
# base_token: base_token
# base_guesses: base_guesses
# base_matches: base_matches
# repeat_count: match[0].length / base_token.length
# lastIndex = j + 1
# matches
end
@doc """
Identifies sequences by looking for repeated differences in unicode codepoint.
This allows skipping, such as 9753, and also matches some extended unicode sequences
such as Greek and Cyrillic alphabets.
For example, consider the input `'abcdb975zy'`
password: a b c d b 9 7 5 z y
index: 0 1 2 3 4 5 6 7 8 9
delta: 1 1 1 -2 -41 -2 -2 69 1
Expected result:
[(i, j, delta), ...] = [(0, 3, 1), (5, 7, -2), (8, 9, 1)]
"""
def sequence_match(_password) do
# return [] if password.length == 1
# update = (i, j, delta) =>
# if j - i > 1 or Math.abs(delta) == 1
# if 0 < Math.abs(delta) <= @MAX_DELTA
# token = password[i..j]
# if /^[a-z]+$/.test(token)
# sequence_name = 'lower'
# sequence_space = 26
# else if /^[A-Z]+$/.test(token)
# sequence_name = 'upper'
# sequence_space = 26
# else if /^\d+$/.test(token)
# sequence_name = 'digits'
# sequence_space = 10
# else
# # conservatively stick with roman alphabet size.
# # (this could be improved)
# sequence_name = 'unicode'
# sequence_space = 26
# result.push
# pattern: 'sequence'
# i: i
# j: j
# token: password[i..j]
# sequence_name: sequence_name
# sequence_space: sequence_space
# ascending: delta > 0
# result = []
# i = 0
# last_delta = null
# for k in [1...password.length]
# delta = password.charCodeAt(k) - password.charCodeAt(k - 1)
# unless last_delta?
# last_delta = delta
# continue if delta == last_delta
# j = k - 1
# update(i, j, last_delta)
# i = j
# last_delta = delta
# update(i, password.length - 1, last_delta)
# result
end
@doc """
Regex matching.
"""
def regex_match(_password, _regexen \\ @regexen) do
# matches = []
# for name, regex of _regexen
# regex.lastIndex = 0 # keeps regex_match stateless
# while rx_match = regex.exec password
# token = rx_match[0]
# matches.push
# pattern: 'regex'
# token: token
# i: rx_match.index
# j: rx_match.index + rx_match[0].length - 1
# regex_name: name
# regex_match: rx_match
# @sorted matches
end
@doc """
A "date" is recognized as:
* any 3-tuple that starts or ends with a 2- or 4-digit year,
* with 2 or 0 separator chars (1.1.91 or 1191),
* maybe zero-padded (01-01-91 vs 1-1-91),
* a month between 1 and 12,
* a day between 1 and 31.
Note: this isn't true date parsing in that "feb 31st" is allowed,
this doesn't check for leap years, etc.
Recipe:
Start with regex to find maybe-dates, then attempt to map the integers
onto month-day-year to filter the maybe-dates into dates.
finally, remove matches that are substrings of other matches to reduce noise.
Note: instead of using a lazy or greedy regex to find many dates over the full string,
this uses a `^...$` regex against every substring of the password -- less performant but leads
to every possible date match.
"""
def date_matching(_password) do
# matches = []
# maybe_date_no_separator = /^\d{4,8}$/
# maybe_date_with_separator = ///
# ^
# ( \d{1,4} ) # day, month, year
# ( [\s/\\_.-] ) # separator
# ( \d{1,2} ) # day, month
# \2 # same separator
# ( \d{1,4} ) # day, month, year
# $
# ///
# # dates without separators are between length 4 '1191' and 8 '11111991'
# for i in [0..password.length - 4]
# for j in [i + 3..i + 7]
# break if j >= password.length
# token = password[i..j]
# continue unless maybe_date_no_separator.exec token
# candidates = []
# for [k,l] in DATE_SPLITS[token.length]
# dmy = @map_ints_to_dmy [
# parseInt token[0...k]
# parseInt token[k...l]
# parseInt token[l...]
# ]
# candidates.push dmy if dmy?
# continue unless candidates.length > 0
# # at this point: different possible dmy mappings for the same i,j substring.
# # match the candidate date that likely takes the fewest guesses: a year closest to 2000.
# # (scoring.REFERENCE_YEAR).
# #
# # ie, considering '111504', prefer 11-15-04 to 1-1-1504
# # (interpreting '04' as 2004)
# best_candidate = candidates[0]
# metric = (candidate) -> Math.abs candidate.year - scoring.REFERENCE_YEAR
# min_distance = metric candidates[0]
# for candidate in candidates[1..]
# distance = metric candidate
# if distance < min_distance
# [best_candidate, min_distance] = [candidate, distance]
# matches.push
# pattern: 'date'
# token: token
# i: i
# j: j
# separator: ''
# year: best_candidate.year
# month: best_candidate.month
# day: best_candidate.day
# # dates with separators are between length 6 '1/1/91' and 10 '11/11/1991'
# for i in [0..password.length - 6]
# for j in [i + 5..i + 9]
# break if j >= password.length
# token = password[i..j]
# rx_match = maybe_date_with_separator.exec token
# continue unless rx_match?
# dmy = @map_ints_to_dmy [
# parseInt rx_match[1]
# parseInt rx_match[3]
# parseInt rx_match[4]
# ]
# continue unless dmy?
# matches.push
# pattern: 'date'
# token: token
# i: i
# j: j
# separator: rx_match[2]
# year: dmy.year
# month: dmy.month
# day: dmy.day
# # matches now contains all valid date strings in a way that is tricky to capture
# # with regexes only. while thorough, it will contain some unintuitive noise:
# #
# # '2015_06_04', in addition to matching 2015_06_04, will also contain
# # 5(!) other date matches: 15_06_04, 5_06_04, ..., even 2015 (matched as 5/1/2020)
# #
# # to reduce noise, remove date matches that are strict substrings of others
# @sorted matches.filter (match) ->
# is_submatch = false
# for other_match in matches
# continue if match is other_match
# if other_match.i <= match.i and other_match.j >= match.j
# is_submatch = true
# break
# not is_submatch
end
# ----------------------------------------------------------------------------
# Helpers
# ----------------------------------------------------------------------------
# makes a pruned copy of l33t_table that only includes password's possible substitutions
defp relevant_l33t_subtable(_password, _table) do
# password_chars = {}
# for chr in password.split('')
# password_chars[chr] = true
# subtable = {}
# for letter, subs of table
# relevant_subs = (sub for sub in subs when sub of password_chars)
# if relevant_subs.length > 0
# subtable[letter] = relevant_subs
# subtable
end
# returns the list of possible 1337 replacement dictionaries for a given password
defp enumerate_l33t_subs(_table) do
# keys = (k for k of table)
# subs = [[]]
# dedup = (subs) ->
# deduped = []
# members = {}
# for sub in subs
# assoc = ([k,v] for k,v in sub)
# assoc.sort()
# label = (k+','+v for k,v in assoc).join('-')
# unless label of members
# members[label] = true
# deduped.push sub
# deduped
# helper = (keys) ->
# return if not keys.length
# first_key = keys[0]
# rest_keys = keys[1..]
# next_subs = []
# for l33t_chr in table[first_key]
# for sub in subs
# dup_l33t_index = -1
# for i in [0...sub.length]
# if sub[i][0] == l33t_chr
# dup_l33t_index = i
# break
# if dup_l33t_index == -1
# sub_extension = sub.concat [[l33t_chr, first_key]]
# next_subs.push sub_extension
# else
# sub_alternative = sub.slice(0)
# sub_alternative.splice(dup_l33t_index, 1)
# sub_alternative.push [l33t_chr, first_key]
# next_subs.push sub
# next_subs.push sub_alternative
# subs = dedup next_subs
# helper(rest_keys)
# helper(keys)
# sub_dicts = [] # convert from assoc lists to dicts
# for sub in subs
# sub_dict = {}
# for [l33t_chr, chr] in sub
# sub_dict[l33t_chr] = chr
# sub_dicts.push sub_dict
# sub_dicts
end
defp spatial_match_helper(_password, _graph, _graph_name) do
# matches = []
# i = 0
# while i < password.length - 1
# j = i + 1
# last_direction = null
# turns = 0
# if graph_name in ['qwerty', 'dvorak'] and @SHIFTED_RX.exec(password.charAt(i))
# # initial character is shifted
# shifted_count = 1
# else
# shifted_count = 0
# loop
# prev_char = password.charAt(j-1)
# found = false
# found_direction = -1
# cur_direction = -1
# adjacents = graph[prev_char] or []
# # consider growing pattern by one character if j hasn't gone over the edge.
# if j < password.length
# cur_char = password.charAt(j)
# for adj in adjacents
# cur_direction += 1
# if adj and adj.indexOf(cur_char) != -1
# found = true
# found_direction = cur_direction
# if adj.indexOf(cur_char) == 1
# # index 1 in the adjacency means the key is shifted,
# # 0 means unshifted: A vs a, % vs 5, etc.
# # for example, 'q' is adjacent to the entry '2@'.
# # @ is shifted w/ index 1, 2 is unshifted.
# shifted_count += 1
# if last_direction != found_direction
# # adding a turn is correct even in the initial case when last_direction is null:
# # every spatial pattern starts with a turn.
# turns += 1
# last_direction = found_direction
# break
# # if the current pattern continued, extend j and try to grow again
# if found
# j += 1
# # otherwise push the pattern discovered so far, if any...
# else
# if j - i > 2 # don't consider length 1 or 2 chains.
# matches.push
# pattern: 'spatial'
# i: i
# j: j-1
# token: password[i...j]
# graph: graph_name
# turns: turns
# shifted_count: shifted_count
# # ...and then start a new search for the rest of the password.
# i = j
# break
# matches
end
# Given a 3-tuple, discard if:
# * middle int is over 31 (for all dmy formats, years are never allowed in the middle)
# * middle int is zero
# * any int is over the max allowable year
# * any int is over two digits but under the min allowable year
# * 2 ints are over 31, the max allowable day
# * 2 ints are zero
# * all ints are over 12, the max allowable month
defp map_ints_to_dmy(_ints) do
# return if ints[1] > 31 or ints[1] <= 0
# over_12 = 0
# over_31 = 0
# under_1 = 0
# for int in ints
# return if 99 < int < DATE_MIN_YEAR or int > DATE_MAX_YEAR
# over_31 += 1 if int > 31
# over_12 += 1 if int > 12
# under_1 += 1 if int <= 0
# return if over_31 >= 2 or over_12 == 3 or under_1 >= 2
# # first look for a four digit year: yyyy + daymonth or daymonth + yyyy
# possible_year_splits = [
# [ints[2], ints[0..1]] # year last
# [ints[0], ints[1..2]] # year first
# ]
# for [y, rest] in possible_year_splits
# if DATE_MIN_YEAR <= y <= DATE_MAX_YEAR
# dm = @map_ints_to_dm rest
# if dm?
# return {
# year: y
# month: dm.month
# day: dm.day
# }
# else
# # for a candidate that includes a four-digit year,
# # when the remaining ints don't match to a day and month,
# # it is not a date.
# return
# # given no four-digit year, two digit years are the most flexible int to match, so
# # try to parse a day-month out of ints[0..1] or ints[1..0]
# for [y, rest] in possible_year_splits
# dm = @map_ints_to_dm rest
# if dm?
# y = @two_to_four_digit_year y
# return {
# year: y
# month: dm.month
# day: dm.day
# }
end
defp map_ints_to_dm(_ints) do
# for [d, m] in [ints, ints.slice().reverse()]
# if 1 <= d <= 31 and 1 <= m <= 12
# return {
# day: d
# month: m
# }
end
defp two_to_four_digit_year(year) do
cond do
year > 99 -> year
# 87 -> 1987
year > 50 -> year + 1900
# 15 -> 2015
true -> year + 2000
end
end
defp ranked_dictionaries do
Enum.map(frequency_lists(), fn {k, v} ->
{k, build_ranked_dict(v)}
end)
end
defp build_ranked_dict(_ordered_list) do
# result = {}
# i = 1 # rank starts at 1, not 0
# for word in ordered_list
# result[word] = i
# i += 1
# result
end
defp empty?(_obj) do
# (k for k of obj).length == 0
end
defp extend(_list1, _list2) do
# lst.push.apply lst, lst2
end
defp translate(_string, _chr_map) do
# (chr_map[chr] or chr for chr in string.split('')).join('')
end
defp mod(_n, _m) do
# ((n % m) + m) % m # mod impl that works for negative numbers
end
def sorted([]), do: []
def sorted(matches) do
matches |> Enum.sort(&(&1.i < &2.i or (&1.i == &2.i and &1.j < &2.j)))
end
defp frequency_lists, do: Zxcvbn.Data.FrequencyLists.all()
defp graphs, do: Zxcvbn.Data.AdjacencyGraphs.all()
end