-
Notifications
You must be signed in to change notification settings - Fork 42
Expand file tree
/
Copy pathEncodingDetector.cs
More file actions
363 lines (315 loc) · 15 KB
/
EncodingDetector.cs
File metadata and controls
363 lines (315 loc) · 15 KB
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
// Copyright Tao Klerks, 2010-2012, tao@klerks.biz
// Licensed under the modified BSD license.
using System;
using System.IO;
using System.Text;
using System.Text.RegularExpressions;
namespace FastColoredTextBoxNS
{
public static class EncodingDetector
{
const long _defaultHeuristicSampleSize = 0x10000; //completely arbitrary - inappropriate for high numbers of files / high speed requirements
public static Encoding DetectTextFileEncoding(string InputFilename)
{
using (FileStream textfileStream = File.OpenRead(InputFilename))
{
return DetectTextFileEncoding(textfileStream, _defaultHeuristicSampleSize);
}
}
public static Encoding DetectTextFileEncoding(FileStream InputFileStream, long HeuristicSampleSize)
{
bool uselessBool = false;
return DetectTextFileEncoding(InputFileStream, _defaultHeuristicSampleSize, out uselessBool);
}
public static Encoding DetectTextFileEncoding(FileStream InputFileStream, long HeuristicSampleSize, out bool HasBOM)
{
Encoding encodingFound = null;
long originalPos = InputFileStream.Position;
InputFileStream.Position = 0;
//First read only what we need for BOM detection
byte[] bomBytes = new byte[InputFileStream.Length > 4 ? 4 : InputFileStream.Length];
InputFileStream.Read(bomBytes, 0, bomBytes.Length);
encodingFound = DetectBOMBytes(bomBytes);
if (encodingFound != null)
{
InputFileStream.Position = originalPos;
HasBOM = true;
return encodingFound;
}
//BOM Detection failed, going for heuristics now.
// create sample byte array and populate it
byte[] sampleBytes = new byte[HeuristicSampleSize > InputFileStream.Length ? InputFileStream.Length : HeuristicSampleSize];
Array.Copy(bomBytes, sampleBytes, bomBytes.Length);
if (InputFileStream.Length > bomBytes.Length)
InputFileStream.Read(sampleBytes, bomBytes.Length, sampleBytes.Length - bomBytes.Length);
InputFileStream.Position = originalPos;
//test byte array content
encodingFound = DetectUnicodeInByteSampleByHeuristics(sampleBytes);
HasBOM = false;
return encodingFound;
}
public static Encoding DetectBOMBytes(byte[] BOMBytes)
{
if (BOMBytes.Length < 2)
return null;
if (BOMBytes[0] == 0xff
&& BOMBytes[1] == 0xfe
&& (BOMBytes.Length < 4
|| BOMBytes[2] != 0
|| BOMBytes[3] != 0
)
)
return Encoding.Unicode;
if (BOMBytes[0] == 0xfe
&& BOMBytes[1] == 0xff
)
return Encoding.BigEndianUnicode;
if (BOMBytes.Length < 3)
return null;
if (BOMBytes[0] == 0xef && BOMBytes[1] == 0xbb && BOMBytes[2] == 0xbf)
return Encoding.UTF8;
if (BOMBytes[0] == 0x2b && BOMBytes[1] == 0x2f && BOMBytes[2] == 0x76)
return Encoding.UTF7;
if (BOMBytes.Length < 4)
return null;
if (BOMBytes[0] == 0xff && BOMBytes[1] == 0xfe && BOMBytes[2] == 0 && BOMBytes[3] == 0)
return Encoding.UTF32;
if (BOMBytes[0] == 0 && BOMBytes[1] == 0 && BOMBytes[2] == 0xfe && BOMBytes[3] == 0xff)
return Encoding.GetEncoding(12001);
return null;
}
public static Encoding DetectUnicodeInByteSampleByHeuristics(byte[] SampleBytes)
{
long oddBinaryNullsInSample = 0;
long evenBinaryNullsInSample = 0;
long suspiciousUTF8SequenceCount = 0;
long suspiciousUTF8BytesTotal = 0;
long likelyUSASCIIBytesInSample = 0;
//Cycle through, keeping count of binary null positions, possible UTF-8
// sequences from upper ranges of Windows-1252, and probable US-ASCII
// character counts.
long currentPos = 0;
int skipUTF8Bytes = 0;
while (currentPos < SampleBytes.Length)
{
//binary null distribution
if (SampleBytes[currentPos] == 0)
{
if (currentPos % 2 == 0)
evenBinaryNullsInSample++;
else
oddBinaryNullsInSample++;
}
//likely US-ASCII characters
if (IsCommonUSASCIIByte(SampleBytes[currentPos]))
likelyUSASCIIBytesInSample++;
//suspicious sequences (look like UTF-8)
if (skipUTF8Bytes == 0)
{
int lengthFound = DetectSuspiciousUTF8SequenceLength(SampleBytes, currentPos);
if (lengthFound > 0)
{
suspiciousUTF8SequenceCount++;
suspiciousUTF8BytesTotal += lengthFound;
skipUTF8Bytes = lengthFound - 1;
}
}
else
{
skipUTF8Bytes--;
}
currentPos++;
}
//1: UTF-16 LE - in english / european environments, this is usually characterized by a
// high proportion of odd binary nulls (starting at 0), with (as this is text) a low
// proportion of even binary nulls.
// The thresholds here used (less than 20% nulls where you expect non-nulls, and more than
// 60% nulls where you do expect nulls) are completely arbitrary.
if (((evenBinaryNullsInSample * 2.0) / SampleBytes.Length) < 0.2
&& ((oddBinaryNullsInSample * 2.0) / SampleBytes.Length) > 0.6
)
return Encoding.Unicode;
//2: UTF-16 BE - in english / european environments, this is usually characterized by a
// high proportion of even binary nulls (starting at 0), with (as this is text) a low
// proportion of odd binary nulls.
// The thresholds here used (less than 20% nulls where you expect non-nulls, and more than
// 60% nulls where you do expect nulls) are completely arbitrary.
if (((oddBinaryNullsInSample * 2.0) / SampleBytes.Length) < 0.2
&& ((evenBinaryNullsInSample * 2.0) / SampleBytes.Length) > 0.6
)
return Encoding.BigEndianUnicode;
//3: UTF-8 - Martin Dürst outlines a method for detecting whether something CAN be UTF-8 content
// using regexp, in his w3c.org unicode FAQ entry:
// http://www.w3.org/International/questions/qa-forms-utf-8
// adapted here for C#.
string potentiallyMangledString = Encoding.ASCII.GetString(SampleBytes);
Regex UTF8Validator = new Regex(@"\A("
+ @"[\x09\x0A\x0D\x20-\x7E]"
+ @"|[\xC2-\xDF][\x80-\xBF]"
+ @"|\xE0[\xA0-\xBF][\x80-\xBF]"
+ @"|[\xE1-\xEC\xEE\xEF][\x80-\xBF]{2}"
+ @"|\xED[\x80-\x9F][\x80-\xBF]"
+ @"|\xF0[\x90-\xBF][\x80-\xBF]{2}"
+ @"|[\xF1-\xF3][\x80-\xBF]{3}"
+ @"|\xF4[\x80-\x8F][\x80-\xBF]{2}"
+ @")*\z");
if (UTF8Validator.IsMatch(potentiallyMangledString))
{
//Unfortunately, just the fact that it CAN be UTF-8 doesn't tell you much about probabilities.
//If all the characters are in the 0-127 range, no harm done, most western charsets are same as UTF-8 in these ranges.
//If some of the characters were in the upper range (western accented characters), however, they would likely be mangled to 2-byte by the UTF-8 encoding process.
// So, we need to play stats.
// The "Random" likelihood of any pair of randomly generated characters being one
// of these "suspicious" character sequences is:
// 128 / (256 * 256) = 0.2%.
//
// In western text data, that is SIGNIFICANTLY reduced - most text data stays in the <127
// character range, so we assume that more than 1 in 500,000 of these character
// sequences indicates UTF-8. The number 500,000 is completely arbitrary - so sue me.
//
// We can only assume these character sequences will be rare if we ALSO assume that this
// IS in fact western text - in which case the bulk of the UTF-8 encoded data (that is
// not already suspicious sequences) should be plain US-ASCII bytes. This, I
// arbitrarily decided, should be 80% (a random distribution, eg binary data, would yield
// approx 40%, so the chances of hitting this threshold by accident in random data are
// VERY low).
if ((suspiciousUTF8SequenceCount * 500000.0 / SampleBytes.Length >= 1) //suspicious sequences
&& (
//all suspicious, so cannot evaluate proportion of US-Ascii
SampleBytes.Length - suspiciousUTF8BytesTotal == 0
||
likelyUSASCIIBytesInSample * 1.0 / (SampleBytes.Length - suspiciousUTF8BytesTotal) >= 0.8
)
)
return Encoding.UTF8;
}
return null;
}
private static bool IsCommonUSASCIIByte(byte testByte)
{
if (testByte == 0x0A //lf
|| testByte == 0x0D //cr
|| testByte == 0x09 //tab
|| (testByte >= 0x20 && testByte <= 0x2F) //common punctuation
|| (testByte >= 0x30 && testByte <= 0x39) //digits
|| (testByte >= 0x3A && testByte <= 0x40) //common punctuation
|| (testByte >= 0x41 && testByte <= 0x5A) //capital letters
|| (testByte >= 0x5B && testByte <= 0x60) //common punctuation
|| (testByte >= 0x61 && testByte <= 0x7A) //lowercase letters
|| (testByte >= 0x7B && testByte <= 0x7E) //common punctuation
)
return true;
else
return false;
}
private static int DetectSuspiciousUTF8SequenceLength(byte[] SampleBytes, long currentPos)
{
int lengthFound = 0;
if (SampleBytes.Length >= currentPos + 1
&& SampleBytes[currentPos] == 0xC2
)
{
if (SampleBytes[currentPos + 1] == 0x81
|| SampleBytes[currentPos + 1] == 0x8D
|| SampleBytes[currentPos + 1] == 0x8F
)
lengthFound = 2;
else if (SampleBytes[currentPos + 1] == 0x90
|| SampleBytes[currentPos + 1] == 0x9D
)
lengthFound = 2;
else if (SampleBytes[currentPos + 1] >= 0xA0
&& SampleBytes[currentPos + 1] <= 0xBF
)
lengthFound = 2;
}
else if (SampleBytes.Length >= currentPos + 1
&& SampleBytes[currentPos] == 0xC3
)
{
if (SampleBytes[currentPos + 1] >= 0x80
&& SampleBytes[currentPos + 1] <= 0xBF
)
lengthFound = 2;
}
else if (SampleBytes.Length >= currentPos + 1
&& SampleBytes[currentPos] == 0xC5
)
{
if (SampleBytes[currentPos + 1] == 0x92
|| SampleBytes[currentPos + 1] == 0x93
)
lengthFound = 2;
else if (SampleBytes[currentPos + 1] == 0xA0
|| SampleBytes[currentPos + 1] == 0xA1
)
lengthFound = 2;
else if (SampleBytes[currentPos + 1] == 0xB8
|| SampleBytes[currentPos + 1] == 0xBD
|| SampleBytes[currentPos + 1] == 0xBE
)
lengthFound = 2;
}
else if (SampleBytes.Length >= currentPos + 1
&& SampleBytes[currentPos] == 0xC6
)
{
if (SampleBytes[currentPos + 1] == 0x92)
lengthFound = 2;
}
else if (SampleBytes.Length >= currentPos + 1
&& SampleBytes[currentPos] == 0xCB
)
{
if (SampleBytes[currentPos + 1] == 0x86
|| SampleBytes[currentPos + 1] == 0x9C
)
lengthFound = 2;
}
else if (SampleBytes.Length >= currentPos + 2
&& SampleBytes[currentPos] == 0xE2
)
{
if (SampleBytes[currentPos + 1] == 0x80)
{
if (SampleBytes[currentPos + 2] == 0x93
|| SampleBytes[currentPos + 2] == 0x94
)
lengthFound = 3;
if (SampleBytes[currentPos + 2] == 0x98
|| SampleBytes[currentPos + 2] == 0x99
|| SampleBytes[currentPos + 2] == 0x9A
)
lengthFound = 3;
if (SampleBytes[currentPos + 2] == 0x9C
|| SampleBytes[currentPos + 2] == 0x9D
|| SampleBytes[currentPos + 2] == 0x9E
)
lengthFound = 3;
if (SampleBytes[currentPos + 2] == 0xA0
|| SampleBytes[currentPos + 2] == 0xA1
|| SampleBytes[currentPos + 2] == 0xA2
)
lengthFound = 3;
if (SampleBytes[currentPos + 2] == 0xA6)
lengthFound = 3;
if (SampleBytes[currentPos + 2] == 0xB0)
lengthFound = 3;
if (SampleBytes[currentPos + 2] == 0xB9
|| SampleBytes[currentPos + 2] == 0xBA
)
lengthFound = 3;
}
else if (SampleBytes[currentPos + 1] == 0x82
&& SampleBytes[currentPos + 2] == 0xAC
)
lengthFound = 3;
else if (SampleBytes[currentPos + 1] == 0x84
&& SampleBytes[currentPos + 2] == 0xA2
)
lengthFound = 3;
}
return lengthFound;
}
}
}