json-repair 0.1.0 → 0.2.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/.rubocop.yml +7 -7
- data/CHANGELOG.md +5 -1
- data/README.md +10 -3
- data/lib/json/repair/version.rb +1 -1
- data/lib/json/repair.rb +4 -6
- data/lib/json/repairer.rb +645 -0
- data/sig/json/repair.rbs +2 -2
- metadata +3 -3
- data/lib/json/repair/repairer.rb +0 -647
@@ -0,0 +1,645 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
require_relative 'repair/string_utils'
|
4
|
+
|
5
|
+
module JSON
|
6
|
+
class Repairer
|
7
|
+
include Repair::StringUtils
|
8
|
+
|
9
|
+
CONTROL_CHARACTERS = {
|
10
|
+
"\b" => '\b',
|
11
|
+
"\f" => '\f',
|
12
|
+
"\n" => '\n',
|
13
|
+
"\r" => '\r',
|
14
|
+
"\t" => '\t'
|
15
|
+
}.freeze
|
16
|
+
|
17
|
+
ESCAPE_CHARACTERS = {
|
18
|
+
'"' => '"',
|
19
|
+
'\\' => '\\',
|
20
|
+
'/' => '/',
|
21
|
+
'b' => "\b",
|
22
|
+
'f' => "\f",
|
23
|
+
'n' => "\n",
|
24
|
+
'r' => "\r",
|
25
|
+
't' => "\t"
|
26
|
+
}.freeze
|
27
|
+
|
28
|
+
def initialize(json)
|
29
|
+
@json = json
|
30
|
+
@index = 0
|
31
|
+
@output = ''
|
32
|
+
end
|
33
|
+
|
34
|
+
def repair
|
35
|
+
processed = parse_value
|
36
|
+
|
37
|
+
throw_unexpected_end unless processed
|
38
|
+
|
39
|
+
processed_comma = parse_character(COMMA)
|
40
|
+
parse_whitespace_and_skip_comments if processed_comma
|
41
|
+
|
42
|
+
if start_of_value?(@json[@index]) && ends_with_comma_or_newline?(@output)
|
43
|
+
# start of a new value after end of the root level object: looks like
|
44
|
+
# newline delimited JSON -> turn into a root level array
|
45
|
+
unless processed_comma
|
46
|
+
# repair missing comma
|
47
|
+
@output = insert_before_last_whitespace(@output, ',')
|
48
|
+
end
|
49
|
+
|
50
|
+
parse_newline_delimited_json
|
51
|
+
elsif processed_comma
|
52
|
+
# repair: remove trailing comma
|
53
|
+
@output = strip_last_occurrence(@output, ',')
|
54
|
+
end
|
55
|
+
|
56
|
+
# repair redundant end quotes
|
57
|
+
while @json[@index] == CLOSING_BRACE || @json[@index] == CLOSING_BRACKET
|
58
|
+
@index += 1
|
59
|
+
parse_whitespace_and_skip_comments
|
60
|
+
end
|
61
|
+
|
62
|
+
if @index >= @json.length
|
63
|
+
# reached the end of the document properly
|
64
|
+
return @output
|
65
|
+
end
|
66
|
+
|
67
|
+
throw_unexpected_character
|
68
|
+
end
|
69
|
+
|
70
|
+
private
|
71
|
+
|
72
|
+
def parse_value
|
73
|
+
parse_whitespace_and_skip_comments
|
74
|
+
process = parse_object || parse_array || parse_string || parse_number || parse_keywords || parse_unquoted_string
|
75
|
+
parse_whitespace_and_skip_comments
|
76
|
+
|
77
|
+
process
|
78
|
+
end
|
79
|
+
|
80
|
+
def parse_whitespace
|
81
|
+
whitespace = ''
|
82
|
+
while @json[@index] && (whitespace?(@json[@index]) || special_whitespace?(@json[@index]))
|
83
|
+
whitespace += whitespace?(@json[@index]) ? @json[@index] : ' '
|
84
|
+
|
85
|
+
@index += 1
|
86
|
+
end
|
87
|
+
|
88
|
+
unless whitespace.empty?
|
89
|
+
@output += whitespace
|
90
|
+
return true
|
91
|
+
end
|
92
|
+
|
93
|
+
false
|
94
|
+
end
|
95
|
+
|
96
|
+
def parse_comment
|
97
|
+
if @json[@index] == '/' && @json[@index + 1] == '*'
|
98
|
+
# Block comment
|
99
|
+
@index += 2
|
100
|
+
@index += 1 until @json[@index].nil? || (@json[@index] == '*' && @json[@index + 1] == '/')
|
101
|
+
@index += 2
|
102
|
+
true
|
103
|
+
elsif @json[@index] == '/' && @json[@index + 1] == '/'
|
104
|
+
# Line comment
|
105
|
+
@index += 2
|
106
|
+
@index += 1 until @json[@index].nil? || @json[@index] == "\n"
|
107
|
+
true
|
108
|
+
else
|
109
|
+
false
|
110
|
+
end
|
111
|
+
end
|
112
|
+
|
113
|
+
# Parse an object like '{"key": "value"}'
|
114
|
+
def parse_object
|
115
|
+
return false unless @json[@index] == OPENING_BRACE
|
116
|
+
|
117
|
+
@output += '{'
|
118
|
+
@index += 1
|
119
|
+
parse_whitespace_and_skip_comments
|
120
|
+
|
121
|
+
# repair: skip leading comma like in {, message: "hi"}
|
122
|
+
parse_whitespace_and_skip_comments if skip_character(COMMA)
|
123
|
+
|
124
|
+
initial = true
|
125
|
+
while @index < @json.length && @json[@index] != CLOSING_BRACE
|
126
|
+
processed_comma = true
|
127
|
+
if initial
|
128
|
+
initial = false
|
129
|
+
else
|
130
|
+
processed_comma = parse_character(COMMA)
|
131
|
+
unless processed_comma
|
132
|
+
# repair missing comma
|
133
|
+
@output = insert_before_last_whitespace(@output, ',')
|
134
|
+
end
|
135
|
+
parse_whitespace_and_skip_comments
|
136
|
+
end
|
137
|
+
|
138
|
+
skip_ellipsis
|
139
|
+
|
140
|
+
processed_key = parse_string || parse_unquoted_string
|
141
|
+
unless processed_key
|
142
|
+
if @json[@index] == CLOSING_BRACE || @json[@index] == OPENING_BRACE ||
|
143
|
+
@json[@index] == CLOSING_BRACKET || @json[@index] == OPENING_BRACKET ||
|
144
|
+
@json[@index].nil?
|
145
|
+
# repair trailing comma
|
146
|
+
@output = strip_last_occurrence(@output, ',')
|
147
|
+
else
|
148
|
+
throw_object_key_expected
|
149
|
+
end
|
150
|
+
break
|
151
|
+
end
|
152
|
+
|
153
|
+
parse_whitespace_and_skip_comments
|
154
|
+
processed_colon = parse_character(COLON)
|
155
|
+
truncated_text = @index >= @json.length
|
156
|
+
unless processed_colon
|
157
|
+
if start_of_value?(@json[@index]) || truncated_text
|
158
|
+
# repair missing colon
|
159
|
+
@output = insert_before_last_whitespace(@output, ':')
|
160
|
+
else
|
161
|
+
throw_colon_expected
|
162
|
+
end
|
163
|
+
end
|
164
|
+
|
165
|
+
processed_value = parse_value
|
166
|
+
unless processed_value
|
167
|
+
if processed_colon || truncated_text
|
168
|
+
# repair missing object value
|
169
|
+
@output += 'null'
|
170
|
+
else
|
171
|
+
throw_colon_expected
|
172
|
+
end
|
173
|
+
end
|
174
|
+
end
|
175
|
+
|
176
|
+
if @json[@index] == CLOSING_BRACE
|
177
|
+
@output += '}'
|
178
|
+
@index += 1
|
179
|
+
else
|
180
|
+
# repair missing end bracket
|
181
|
+
@output = insert_before_last_whitespace(@output, '}')
|
182
|
+
end
|
183
|
+
|
184
|
+
true
|
185
|
+
end
|
186
|
+
|
187
|
+
def skip_character(char)
|
188
|
+
if @json[@index] == char
|
189
|
+
@index += 1
|
190
|
+
true
|
191
|
+
else
|
192
|
+
false
|
193
|
+
end
|
194
|
+
end
|
195
|
+
|
196
|
+
# Skip ellipsis like "[1,2,3,...]" or "[1,2,3,...,9]" or "[...,7,8,9]"
|
197
|
+
# or a similar construct in objects.
|
198
|
+
def skip_ellipsis
|
199
|
+
parse_whitespace_and_skip_comments
|
200
|
+
|
201
|
+
if @json[@index] == DOT &&
|
202
|
+
@json[@index + 1] == DOT &&
|
203
|
+
@json[@index + 2] == DOT
|
204
|
+
# repair: remove the ellipsis (three dots) and optionally a comma
|
205
|
+
@index += 3
|
206
|
+
parse_whitespace_and_skip_comments
|
207
|
+
skip_character(COMMA)
|
208
|
+
end
|
209
|
+
end
|
210
|
+
|
211
|
+
# Parse a string enclosed by double quotes "...". Can contain escaped quotes
|
212
|
+
# Repair strings enclosed in single quotes or special quotes
|
213
|
+
# Repair an escaped string
|
214
|
+
#
|
215
|
+
# The function can run in two stages:
|
216
|
+
# - First, it assumes the string has a valid end quote
|
217
|
+
# - If it turns out that the string does not have a valid end quote followed
|
218
|
+
# by a delimiter (which should be the case), the function runs again in a
|
219
|
+
# more conservative way, stopping the string at the first next delimiter
|
220
|
+
# and fixing the string by inserting a quote there.
|
221
|
+
def parse_string(stop_at_delimiter: false)
|
222
|
+
if @json[@index] == BACKSLASH
|
223
|
+
# repair: remove the first escape character
|
224
|
+
@index += 1
|
225
|
+
skip_escape_chars = true
|
226
|
+
end
|
227
|
+
|
228
|
+
if quote?(@json[@index])
|
229
|
+
# double quotes are correct JSON,
|
230
|
+
# single quotes come from JavaScript for example, we assume it will have a correct single end quote too
|
231
|
+
# otherwise, we will match any double-quote-like start with a double-quote-like end,
|
232
|
+
# or any single-quote-like start with a single-quote-like end
|
233
|
+
is_end_quote = if double_quote?(@json[@index])
|
234
|
+
method(:double_quote?)
|
235
|
+
elsif single_quote?(@json[@index])
|
236
|
+
method(:single_quote?)
|
237
|
+
elsif single_quote_like?(@json[@index])
|
238
|
+
method(:single_quote_like?)
|
239
|
+
else
|
240
|
+
method(:double_quote_like?)
|
241
|
+
end
|
242
|
+
|
243
|
+
i_before = @index
|
244
|
+
o_before = @output.length
|
245
|
+
|
246
|
+
str = '"'
|
247
|
+
@index += 1
|
248
|
+
|
249
|
+
loop do
|
250
|
+
if @index >= @json.length
|
251
|
+
# end of text, we are missing an end quote
|
252
|
+
|
253
|
+
i_prev = prev_non_whitespace_index(@index - 1)
|
254
|
+
if !stop_at_delimiter && delimiter?(@json[i_prev])
|
255
|
+
# if the text ends with a delimiter, like ["hello],
|
256
|
+
# so the missing end quote should be inserted before this delimiter
|
257
|
+
# retry parsing the string, stopping at the first next delimiter
|
258
|
+
@index = i_before
|
259
|
+
@output = @output[0...o_before]
|
260
|
+
|
261
|
+
return parse_string(stop_at_delimiter: true)
|
262
|
+
end
|
263
|
+
|
264
|
+
# repair missing quote
|
265
|
+
str = insert_before_last_whitespace(str, '"')
|
266
|
+
@output += str
|
267
|
+
|
268
|
+
return true
|
269
|
+
elsif is_end_quote.call(@json[@index])
|
270
|
+
# end quote
|
271
|
+
i_quote = @index
|
272
|
+
o_quote = str.length
|
273
|
+
str += '"'
|
274
|
+
@index += 1
|
275
|
+
@output += str
|
276
|
+
|
277
|
+
parse_whitespace_and_skip_comments
|
278
|
+
|
279
|
+
if stop_at_delimiter ||
|
280
|
+
@index >= @json.length ||
|
281
|
+
delimiter?(@json[@index]) ||
|
282
|
+
quote?(@json[@index]) ||
|
283
|
+
digit?(@json[@index])
|
284
|
+
# The quote is followed by the end of the text, a delimiter, or a next value
|
285
|
+
parse_concatenated_string
|
286
|
+
|
287
|
+
return true
|
288
|
+
end
|
289
|
+
|
290
|
+
if delimiter?(@json[prev_non_whitespace_index(i_quote - 1)])
|
291
|
+
# This is not the right end quote: it is preceded by a delimiter,
|
292
|
+
# and NOT followed by a delimiter. So, there is an end quote missing
|
293
|
+
# parse the string again and then stop at the first next delimiter
|
294
|
+
@index = i_before
|
295
|
+
@output = @output[...o_before]
|
296
|
+
|
297
|
+
return parse_string(stop_at_delimiter: true)
|
298
|
+
end
|
299
|
+
|
300
|
+
# revert to right after the quote but before any whitespace, and continue parsing the string
|
301
|
+
@output = @output[...o_before]
|
302
|
+
@index = i_quote + 1
|
303
|
+
|
304
|
+
# repair unescaped quote
|
305
|
+
str = "#{str[...o_quote]}\\#{str[o_quote..]}"
|
306
|
+
elsif stop_at_delimiter && delimiter?(@json[@index])
|
307
|
+
# we're in the mode to stop the string at the first delimiter
|
308
|
+
# because there is an end quote missing
|
309
|
+
|
310
|
+
# repair missing quote
|
311
|
+
str = insert_before_last_whitespace(str, '"')
|
312
|
+
@output += str
|
313
|
+
|
314
|
+
parse_concatenated_string
|
315
|
+
|
316
|
+
return true
|
317
|
+
elsif @json[@index] == BACKSLASH
|
318
|
+
# handle escaped content like \n or \u2605
|
319
|
+
char = @json[@index + 1]
|
320
|
+
escape_char = ESCAPE_CHARACTERS[char]
|
321
|
+
if escape_char
|
322
|
+
str += @json[@index, 2]
|
323
|
+
@index += 2
|
324
|
+
elsif char == 'u'
|
325
|
+
j = 2
|
326
|
+
j += 1 while j < 6 && @json[@index + j] && hex?(@json[@index + j])
|
327
|
+
if j == 6
|
328
|
+
str += @json[@index, 6]
|
329
|
+
@index += 6
|
330
|
+
elsif @index + j >= @json.length
|
331
|
+
# repair invalid or truncated unicode char at the end of the text
|
332
|
+
# by removing the unicode char and ending the string here
|
333
|
+
@index = @json.length
|
334
|
+
else
|
335
|
+
throw_invalid_unicode_character
|
336
|
+
end
|
337
|
+
else
|
338
|
+
# repair invalid escape character: remove it
|
339
|
+
str += char
|
340
|
+
@index += 2
|
341
|
+
end
|
342
|
+
else
|
343
|
+
# handle regular characters
|
344
|
+
char = @json[@index]
|
345
|
+
|
346
|
+
if char == DOUBLE_QUOTE && @json[@index - 1] != BACKSLASH
|
347
|
+
# repair unescaped double quote
|
348
|
+
str += "\\#{char}"
|
349
|
+
elsif control_character?(char)
|
350
|
+
# unescaped control character
|
351
|
+
str += CONTROL_CHARACTERS[char]
|
352
|
+
else
|
353
|
+
throw_invalid_character(char) unless valid_string_character?(char)
|
354
|
+
str += char
|
355
|
+
end
|
356
|
+
|
357
|
+
@index += 1
|
358
|
+
end
|
359
|
+
|
360
|
+
if skip_escape_chars
|
361
|
+
# repair: skipped escape character (nothing to do)
|
362
|
+
skip_escape_character
|
363
|
+
end
|
364
|
+
end
|
365
|
+
end
|
366
|
+
|
367
|
+
false
|
368
|
+
end
|
369
|
+
|
370
|
+
# Repair an unquoted string by adding quotes around it
|
371
|
+
# Repair a MongoDB function call like NumberLong("2")
|
372
|
+
# Repair a JSONP function call like callback({...});
|
373
|
+
def parse_unquoted_string
|
374
|
+
start = @index
|
375
|
+
@index += 1 while @index < @json.length && !delimiter_except_slash?(@json[@index]) && !quote?(@json[@index])
|
376
|
+
return if @index <= start
|
377
|
+
|
378
|
+
if @json[@index] == '(' && function_name?(@json[start...@index].strip)
|
379
|
+
# Repair a MongoDB function call like NumberLong("2")
|
380
|
+
# Repair a JSONP function call like callback({...});
|
381
|
+
@index += 1
|
382
|
+
|
383
|
+
parse_value
|
384
|
+
|
385
|
+
if @json[@index] == ')'
|
386
|
+
# Repair: skip close bracket of function call
|
387
|
+
@index += 1
|
388
|
+
# Repair: skip semicolon after JSONP call
|
389
|
+
@index += 1 if @json[@index] == ';'
|
390
|
+
end
|
391
|
+
else
|
392
|
+
# Repair unquoted string
|
393
|
+
# Also, repair undefined into null
|
394
|
+
|
395
|
+
# First, go back to prevent getting trailing whitespaces in the string
|
396
|
+
@index -= 1 while whitespace?(@json[@index - 1]) && @index.positive?
|
397
|
+
|
398
|
+
symbol = @json[start...@index]
|
399
|
+
@output += symbol == 'undefined' ? 'null' : symbol.inspect
|
400
|
+
|
401
|
+
if @json[@index] == '"'
|
402
|
+
# We had a missing start quote, but now we encountered the end quote, so we can skip that one
|
403
|
+
@index += 1
|
404
|
+
end
|
405
|
+
end
|
406
|
+
|
407
|
+
true
|
408
|
+
end
|
409
|
+
|
410
|
+
def parse_character(char)
|
411
|
+
if @json[@index] == char
|
412
|
+
@output += @json[@index]
|
413
|
+
@index += 1
|
414
|
+
true
|
415
|
+
else
|
416
|
+
false
|
417
|
+
end
|
418
|
+
end
|
419
|
+
|
420
|
+
def parse_whitespace_and_skip_comments
|
421
|
+
start = @index
|
422
|
+
|
423
|
+
changed = parse_whitespace
|
424
|
+
loop do
|
425
|
+
changed = parse_comment
|
426
|
+
changed = parse_whitespace if changed
|
427
|
+
break unless changed
|
428
|
+
end
|
429
|
+
|
430
|
+
@index > start
|
431
|
+
end
|
432
|
+
|
433
|
+
# Parse a number like 2.4 or 2.4e6
|
434
|
+
def parse_number
|
435
|
+
start = @index
|
436
|
+
if @json[@index] == '-'
|
437
|
+
@index += 1
|
438
|
+
if at_end_of_number?
|
439
|
+
repair_number_ending_with_numeric_symbol(start)
|
440
|
+
return true
|
441
|
+
end
|
442
|
+
unless digit?(@json[@index])
|
443
|
+
@index = start
|
444
|
+
return false
|
445
|
+
end
|
446
|
+
end
|
447
|
+
|
448
|
+
# Note that in JSON leading zeros like "00789" are not allowed.
|
449
|
+
# We will allow all leading zeros here though and at the end of parse_number
|
450
|
+
# check against trailing zeros and repair that if needed.
|
451
|
+
# Leading zeros can have meaning, so we should not clear them.
|
452
|
+
@index += 1 while digit?(@json[@index])
|
453
|
+
|
454
|
+
if @json[@index] == '.'
|
455
|
+
@index += 1
|
456
|
+
if at_end_of_number?
|
457
|
+
repair_number_ending_with_numeric_symbol(start)
|
458
|
+
return true
|
459
|
+
end
|
460
|
+
unless digit?(@json[@index])
|
461
|
+
@index = start
|
462
|
+
return false
|
463
|
+
end
|
464
|
+
@index += 1 while digit?(@json[@index])
|
465
|
+
end
|
466
|
+
|
467
|
+
if @json[@index] && @json[@index].downcase == 'e'
|
468
|
+
@index += 1
|
469
|
+
@index += 1 if ['-', '+'].include?(@json[@index])
|
470
|
+
if at_end_of_number?
|
471
|
+
repair_number_ending_with_numeric_symbol(start)
|
472
|
+
return true
|
473
|
+
end
|
474
|
+
unless digit?(@json[@index])
|
475
|
+
@index = start
|
476
|
+
return false
|
477
|
+
end
|
478
|
+
@index += 1 while digit?(@json[@index])
|
479
|
+
end
|
480
|
+
|
481
|
+
# if we're not at the end of the number by this point, allow this to be parsed as another type
|
482
|
+
unless at_end_of_number?
|
483
|
+
@index = start
|
484
|
+
return false
|
485
|
+
end
|
486
|
+
|
487
|
+
if @index > start
|
488
|
+
# repair a number with leading zeros like "00789"
|
489
|
+
num = @json[start...@index]
|
490
|
+
has_invalid_leading_zero = num.match?(/^0\d/)
|
491
|
+
|
492
|
+
@output += has_invalid_leading_zero ? "\"#{num}\"" : num
|
493
|
+
return true
|
494
|
+
end
|
495
|
+
|
496
|
+
false
|
497
|
+
end
|
498
|
+
|
499
|
+
def at_end_of_number?
|
500
|
+
@index >= @json.length || delimiter?(@json[@index]) || whitespace?(@json[@index])
|
501
|
+
end
|
502
|
+
|
503
|
+
# Parse an array like '["item1", "item2", ...]'
|
504
|
+
def parse_array
|
505
|
+
if @json[@index] == OPENING_BRACKET
|
506
|
+
@output += '['
|
507
|
+
@index += 1
|
508
|
+
parse_whitespace_and_skip_comments
|
509
|
+
|
510
|
+
# repair: skip leading comma like in [,1,2,3]
|
511
|
+
parse_whitespace_and_skip_comments if skip_character(COMMA)
|
512
|
+
|
513
|
+
initial = true
|
514
|
+
while @index < @json.length && @json[@index] != CLOSING_BRACKET
|
515
|
+
if initial
|
516
|
+
initial = false
|
517
|
+
else
|
518
|
+
processed_comma = parse_character(COMMA)
|
519
|
+
# repair missing comma
|
520
|
+
@output = insert_before_last_whitespace(@output, ',') unless processed_comma
|
521
|
+
end
|
522
|
+
|
523
|
+
skip_ellipsis
|
524
|
+
|
525
|
+
processed_value = parse_value
|
526
|
+
next if processed_value
|
527
|
+
|
528
|
+
# repair trailing comma
|
529
|
+
@output = strip_last_occurrence(@output, ',')
|
530
|
+
break
|
531
|
+
end
|
532
|
+
|
533
|
+
if @json[@index] == CLOSING_BRACKET
|
534
|
+
@output += ']'
|
535
|
+
@index += 1
|
536
|
+
else
|
537
|
+
# repair missing closing array bracket
|
538
|
+
@output = insert_before_last_whitespace(@output, ']')
|
539
|
+
end
|
540
|
+
|
541
|
+
true
|
542
|
+
else
|
543
|
+
false
|
544
|
+
end
|
545
|
+
end
|
546
|
+
|
547
|
+
def prev_non_whitespace_index(start)
|
548
|
+
prev = start
|
549
|
+
prev -= 1 while prev.positive? && whitespace?(@json[prev])
|
550
|
+
prev
|
551
|
+
end
|
552
|
+
|
553
|
+
# Repair concatenated strings like "hello" + "world", change this into "helloworld"
|
554
|
+
def parse_concatenated_string
|
555
|
+
processed = false
|
556
|
+
|
557
|
+
parse_whitespace_and_skip_comments
|
558
|
+
while @json[@index] == PLUS
|
559
|
+
processed = true
|
560
|
+
@index += 1
|
561
|
+
parse_whitespace_and_skip_comments
|
562
|
+
|
563
|
+
# repair: remove the end quote of the first string
|
564
|
+
@output = strip_last_occurrence(@output, '"', strip_remaining_text: true)
|
565
|
+
start = @output.length
|
566
|
+
parsed_str = parse_string
|
567
|
+
@output = if parsed_str
|
568
|
+
# repair: remove the start quote of the second string
|
569
|
+
remove_at_index(@output, start, 1)
|
570
|
+
else
|
571
|
+
# repair: remove the '+' because it is not followed by a string
|
572
|
+
insert_before_last_whitespace(@output, '"')
|
573
|
+
end
|
574
|
+
end
|
575
|
+
|
576
|
+
processed
|
577
|
+
end
|
578
|
+
|
579
|
+
def repair_number_ending_with_numeric_symbol(start)
|
580
|
+
# repair numbers cut off at the end
|
581
|
+
# this will only be called when we end after a '.', '-', or 'e' and does not
|
582
|
+
# change the number more than it needs to make it valid JSON
|
583
|
+
@output += "#{@json[start...@index]}0"
|
584
|
+
end
|
585
|
+
|
586
|
+
# Parse and repair Newline Delimited JSON (NDJSON):
|
587
|
+
# multiple JSON objects separated by a newline character
|
588
|
+
def parse_newline_delimited_json
|
589
|
+
# repair NDJSON
|
590
|
+
initial = true
|
591
|
+
processed_value = true
|
592
|
+
while processed_value
|
593
|
+
if initial
|
594
|
+
initial = false
|
595
|
+
else
|
596
|
+
# parse optional comma, insert when missing
|
597
|
+
processed_comma = parse_character(COMMA)
|
598
|
+
unless processed_comma
|
599
|
+
# repair: add missing comma
|
600
|
+
@output = insert_before_last_whitespace(@output, ',')
|
601
|
+
end
|
602
|
+
end
|
603
|
+
|
604
|
+
processed_value = parse_value
|
605
|
+
end
|
606
|
+
|
607
|
+
unless processed_value
|
608
|
+
# repair: remove trailing comma
|
609
|
+
@output = strip_last_occurrence(@output, ',')
|
610
|
+
end
|
611
|
+
|
612
|
+
# repair: wrap the output inside array brackets
|
613
|
+
@output = "[\n#{@output}\n]"
|
614
|
+
end
|
615
|
+
|
616
|
+
def skip_escape_character
|
617
|
+
skip_character(BACKSLASH)
|
618
|
+
end
|
619
|
+
|
620
|
+
def throw_invalid_character(char)
|
621
|
+
raise JSONRepairError, "Invalid character #{char.inspect} at index #{@index}"
|
622
|
+
end
|
623
|
+
|
624
|
+
def throw_unexpected_character
|
625
|
+
raise JSONRepairError, "Unexpected character #{@json[@index].inspect} at index #{@index}"
|
626
|
+
end
|
627
|
+
|
628
|
+
def throw_unexpected_end
|
629
|
+
raise JSONRepairError, 'Unexpected end of json string'
|
630
|
+
end
|
631
|
+
|
632
|
+
def throw_object_key_expected
|
633
|
+
raise JSONRepairError, 'Object key expected'
|
634
|
+
end
|
635
|
+
|
636
|
+
def throw_colon_expected
|
637
|
+
raise JSONRepairError, 'Colon expected'
|
638
|
+
end
|
639
|
+
|
640
|
+
def throw_invalid_unicode_character
|
641
|
+
chars = @json[@index, 6]
|
642
|
+
raise JSONRepairError, "Invalid unicode character #{chars.inspect} at index #{@index}"
|
643
|
+
end
|
644
|
+
end
|
645
|
+
end
|
data/sig/json/repair.rbs
CHANGED
metadata
CHANGED
@@ -1,14 +1,14 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: json-repair
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 0.
|
4
|
+
version: 0.2.0
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- Aleksandr Zykov
|
8
8
|
autorequire:
|
9
9
|
bindir: exe
|
10
10
|
cert_chain: []
|
11
|
-
date: 2024-
|
11
|
+
date: 2024-06-04 00:00:00.000000000 Z
|
12
12
|
dependencies: []
|
13
13
|
description: This is a simple gem that repairs broken JSON strings.
|
14
14
|
email:
|
@@ -25,9 +25,9 @@ files:
|
|
25
25
|
- README.md
|
26
26
|
- Rakefile
|
27
27
|
- lib/json/repair.rb
|
28
|
-
- lib/json/repair/repairer.rb
|
29
28
|
- lib/json/repair/string_utils.rb
|
30
29
|
- lib/json/repair/version.rb
|
30
|
+
- lib/json/repairer.rb
|
31
31
|
- sig/json/repair.rbs
|
32
32
|
homepage: https://github.com/sashazykov/json-repair-rb
|
33
33
|
licenses:
|