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
import logging
from os.path import basename, splitext
from typing import BinaryIO, List, Optional, Set

try:
    from os import PathLike
except ImportError:  # pragma: no cover
    PathLike = str  # type: ignore

from .cd import (
    coherence_ratio,
    encoding_languages,
    mb_encoding_languages,
    merge_coherence_ratios,
)
from .constant import IANA_SUPPORTED, TOO_BIG_SEQUENCE, TOO_SMALL_SEQUENCE, TRACE
from .md import mess_ratio
from .models import CharsetMatch, CharsetMatches
from .utils import (
    any_specified_encoding,
    iana_name,
    identify_sig_or_bom,
    is_cp_similar,
    is_multi_byte_encoding,
    should_strip_sig_or_bom,
)

# Will most likely be controversial
# logging.addLevelName(TRACE, "TRACE")
logger = logging.getLogger("charset_normalizer")
explain_handler = logging.StreamHandler()
explain_handler.setFormatter(
    logging.Formatter("%(asctime)s | %(levelname)s | %(message)s")
)


def from_bytes(
    sequences: bytes,
    steps: int = 5,
    chunk_size: int = 512,
    threshold: float = 0.2,
    cp_isolation: List[str] = None,
    cp_exclusion: List[str] = None,
    preemptive_behaviour: bool = True,
    explain: bool = False,
) -> CharsetMatches:
    """
    Given a raw bytes sequence, return the best possibles charset usable to render str objects.
    If there is no results, it is a strong indicator that the source is binary/not text.
    By default, the process will extract 5 blocs of 512o each to assess the mess and coherence of a given sequence.
    And will give up a particular code page after 20% of measured mess. Those criteria are customizable at will.

    The preemptive behavior DOES NOT replace the traditional detection workflow, it prioritize a particular code page
    but never take it for granted. Can improve the performance.

    You may want to focus your attention to some code page or/and not others, use cp_isolation and cp_exclusion for that
    purpose.

    This function will strip the SIG in the payload/sequence every time except on UTF-16, UTF-32.
    By default the library does not setup any handler other than the NullHandler, if you choose to set the 'explain'
    toggle to True it will alter the logger configuration to add a StreamHandler that is suitable for debugging.
    Custom logging format and handler can be set manually.
    """

    if not isinstance(sequences, (bytearray, bytes)):
        raise TypeError(
            "Expected object of type bytes or bytearray, got: {0}".format(
                type(sequences)
            )
        )

    if explain:
        previous_logger_level = logger.level  # type: int
        logger.addHandler(explain_handler)
        logger.setLevel(TRACE)

    length = len(sequences)  # type: int

    if length == 0:
        logger.debug("Encoding detection on empty bytes, assuming utf_8 intention.")
        if explain:
            logger.removeHandler(explain_handler)
            logger.setLevel(previous_logger_level or logging.WARNING)
        return CharsetMatches([CharsetMatch(sequences, "utf_8", 0.0, False, [], "")])

    if cp_isolation is not None:
        logger.log(
            TRACE,
            "cp_isolation is set. use this flag for debugging purpose. "
            "limited list of encoding allowed : %s.",
            ", ".join(cp_isolation),
        )
        cp_isolation = [iana_name(cp, False) for cp in cp_isolation]
    else:
        cp_isolation = []

    if cp_exclusion is not None:
        logger.log(
            TRACE,
            "cp_exclusion is set. use this flag for debugging purpose. "
            "limited list of encoding excluded : %s.",
            ", ".join(cp_exclusion),
        )
        cp_exclusion = [iana_name(cp, False) for cp in cp_exclusion]
    else:
        cp_exclusion = []

    if length <= (chunk_size * steps):
        logger.log(
            TRACE,
            "override steps (%i) and chunk_size (%i) as content does not fit (%i byte(s) given) parameters.",
            steps,
            chunk_size,
            length,
        )
        steps = 1
        chunk_size = length

    if steps > 1 and length / steps < chunk_size:
        chunk_size = int(length / steps)

    is_too_small_sequence = len(sequences) < TOO_SMALL_SEQUENCE  # type: bool
    is_too_large_sequence = len(sequences) >= TOO_BIG_SEQUENCE  # type: bool

    if is_too_small_sequence:
        logger.log(
            TRACE,
            "Trying to detect encoding from a tiny portion of ({}) byte(s).".format(
                length
            ),
        )
    elif is_too_large_sequence:
        logger.log(
            TRACE,
            "Using lazy str decoding because the payload is quite large, ({}) byte(s).".format(
                length
            ),
        )

    prioritized_encodings = []  # type: List[str]

    specified_encoding = (
        any_specified_encoding(sequences) if preemptive_behaviour else None
    )  # type: Optional[str]

    if specified_encoding is not None:
        prioritized_encodings.append(specified_encoding)
        logger.log(
            TRACE,
            "Detected declarative mark in sequence. Priority +1 given for %s.",
            specified_encoding,
        )

    tested = set()  # type: Set[str]
    tested_but_hard_failure = []  # type: List[str]
    tested_but_soft_failure = []  # type: List[str]

    fallback_ascii = None  # type: Optional[CharsetMatch]
    fallback_u8 = None  # type: Optional[CharsetMatch]
    fallback_specified = None  # type: Optional[CharsetMatch]

    results = CharsetMatches()  # type: CharsetMatches

    sig_encoding, sig_payload = identify_sig_or_bom(sequences)

    if sig_encoding is not None:
        prioritized_encodings.append(sig_encoding)
        logger.log(
            TRACE,
            "Detected a SIG or BOM mark on first %i byte(s). Priority +1 given for %s.",
            len(sig_payload),
            sig_encoding,
        )

    prioritized_encodings.append("ascii")

    if "utf_8" not in prioritized_encodings:
        prioritized_encodings.append("utf_8")

    for encoding_iana in prioritized_encodings + IANA_SUPPORTED:

        if cp_isolation and encoding_iana not in cp_isolation:
            continue

        if cp_exclusion and encoding_iana in cp_exclusion:
            continue

        if encoding_iana in tested:
            continue

        tested.add(encoding_iana)

        decoded_payload = None  # type: Optional[str]
        bom_or_sig_available = sig_encoding == encoding_iana  # type: bool
        strip_sig_or_bom = bom_or_sig_available and should_strip_sig_or_bom(
            encoding_iana
        )  # type: bool

        if encoding_iana in {"utf_16", "utf_32"} and not bom_or_sig_available:
            logger.log(
                TRACE,
                "Encoding %s wont be tested as-is because it require a BOM. Will try some sub-encoder LE/BE.",
                encoding_iana,
            )
            continue

        try:
            is_multi_byte_decoder = is_multi_byte_encoding(encoding_iana)  # type: bool
        except (ModuleNotFoundError, ImportError):
            logger.log(
                TRACE,
                "Encoding %s does not provide an IncrementalDecoder",
                encoding_iana,
            )
            continue

        try:
            if is_too_large_sequence and is_multi_byte_decoder is False:
                str(
                    sequences[: int(50e4)]
                    if strip_sig_or_bom is False
                    else sequences[len(sig_payload) : int(50e4)],
                    encoding=encoding_iana,
                )
            else:
                decoded_payload = str(
                    sequences
                    if strip_sig_or_bom is False
                    else sequences[len(sig_payload) :],
                    encoding=encoding_iana,
                )
        except (UnicodeDecodeError, LookupError) as e:
            if not isinstance(e, LookupError):
                logger.log(
                    TRACE,
                    "Code page %s does not fit given bytes sequence at ALL. %s",
                    encoding_iana,
                    str(e),
                )
            tested_but_hard_failure.append(encoding_iana)
            continue

        similar_soft_failure_test = False  # type: bool

        for encoding_soft_failed in tested_but_soft_failure:
            if is_cp_similar(encoding_iana, encoding_soft_failed):
                similar_soft_failure_test = True
                break

        if similar_soft_failure_test:
            logger.log(
                TRACE,
                "%s is deemed too similar to code page %s and was consider unsuited already. Continuing!",
                encoding_iana,
                encoding_soft_failed,
            )
            continue

        r_ = range(
            0 if not bom_or_sig_available else len(sig_payload),
            length,
            int(length / steps),
        )

        multi_byte_bonus = (
            is_multi_byte_decoder
            and decoded_payload is not None
            and len(decoded_payload) < length
        )  # type: bool

        if multi_byte_bonus:
            logger.log(
                TRACE,
                "Code page %s is a multi byte encoding table and it appear that at least one character "
                "was encoded using n-bytes.",
                encoding_iana,
            )

        max_chunk_gave_up = int(len(r_) / 4)  # type: int

        max_chunk_gave_up = max(max_chunk_gave_up, 2)
        early_stop_count = 0  # type: int
        lazy_str_hard_failure = False

        md_chunks = []  # type: List[str]
        md_ratios = []

        for i in r_:
            if i + chunk_size > length + 8:
                continue

            cut_sequence = sequences[i : i + chunk_size]

            if bom_or_sig_available and strip_sig_or_bom is False:
                cut_sequence = sig_payload + cut_sequence

            try:
                chunk = cut_sequence.decode(
                    encoding_iana,
                    errors="ignore" if is_multi_byte_decoder else "strict",
                )  # type: str
            except UnicodeDecodeError as e:  # Lazy str loading may have missed something there
                logger.log(
                    TRACE,
                    "LazyStr Loading: After MD chunk decode, code page %s does not fit given bytes sequence at ALL. %s",
                    encoding_iana,
                    str(e),
                )
                early_stop_count = max_chunk_gave_up
                lazy_str_hard_failure = True
                break

            # multi-byte bad cutting detector and adjustment
            # not the cleanest way to perform that fix but clever enough for now.
            if is_multi_byte_decoder and i > 0 and sequences[i] >= 0x80:

                chunk_partial_size_chk = min(chunk_size, 16)  # type: int

                if (
                    decoded_payload
                    and chunk[:chunk_partial_size_chk] not in decoded_payload
                ):
                    for j in range(i, i - 4, -1):
                        cut_sequence = sequences[j : i + chunk_size]

                        if bom_or_sig_available and strip_sig_or_bom is False:
                            cut_sequence = sig_payload + cut_sequence

                        chunk = cut_sequence.decode(encoding_iana, errors="ignore")

                        if chunk[:chunk_partial_size_chk] in decoded_payload:
                            break

            md_chunks.append(chunk)

            md_ratios.append(mess_ratio(chunk, threshold))

            if md_ratios[-1] >= threshold:
                early_stop_count += 1

            if (early_stop_count >= max_chunk_gave_up) or (
                bom_or_sig_available and strip_sig_or_bom is False
            ):
                break

        # We might want to check the sequence again with the whole content
        # Only if initial MD tests passes
        if (
            not lazy_str_hard_failure
            and is_too_large_sequence
            and not is_multi_byte_decoder
        ):
            try:
                sequences[int(50e3) :].decode(encoding_iana, errors="strict")
            except UnicodeDecodeError as e:
                logger.log(
                    TRACE,
                    "LazyStr Loading: After final lookup, code page %s does not fit given bytes sequence at ALL. %s",
                    encoding_iana,
                    str(e),
                )
                tested_but_hard_failure.append(encoding_iana)
                continue

        mean_mess_ratio = (
            sum(md_ratios) / len(md_ratios) if md_ratios else 0.0
        )  # type: float
        if mean_mess_ratio >= threshold or early_stop_count >= max_chunk_gave_up:
            tested_but_soft_failure.append(encoding_iana)
            logger.log(
                TRACE,
                "%s was excluded because of initial chaos probing. Gave up %i time(s). "
                "Computed mean chaos is %f %%.",
                encoding_iana,
                early_stop_count,
                round(mean_mess_ratio * 100, ndigits=3),
            )
            # Preparing those fallbacks in case we got nothing.
            if (
                encoding_iana in ["ascii", "utf_8", specified_encoding]
                and not lazy_str_hard_failure
            ):
                fallback_entry = CharsetMatch(
                    sequences, encoding_iana, threshold, False, [], decoded_payload
                )
                if encoding_iana == specified_encoding:
                    fallback_specified = fallback_entry
                elif encoding_iana == "ascii":
                    fallback_ascii = fallback_entry
                else:
                    fallback_u8 = fallback_entry
            continue

        logger.log(
            TRACE,
            "%s passed initial chaos probing. Mean measured chaos is %f %%",
            encoding_iana,
            round(mean_mess_ratio * 100, ndigits=3),
        )

        if not is_multi_byte_decoder:
            target_languages = encoding_languages(encoding_iana)  # type: List[str]
        else:
            target_languages = mb_encoding_languages(encoding_iana)

        if target_languages:
            logger.log(
                TRACE,
                "{} should target any language(s) of {}".format(
                    encoding_iana, str(target_languages)
                ),
            )

        cd_ratios = []

        # We shall skip the CD when its about ASCII
        # Most of the time its not relevant to run "language-detection" on it.
        if encoding_iana != "ascii":
            for chunk in md_chunks:
                chunk_languages = coherence_ratio(
                    chunk, 0.1, ",".join(target_languages) if target_languages else None
                )

                cd_ratios.append(chunk_languages)

        cd_ratios_merged = merge_coherence_ratios(cd_ratios)

        if cd_ratios_merged:
            logger.log(
                TRACE,
                "We detected language {} using {}".format(
                    cd_ratios_merged, encoding_iana
                ),
            )

        results.append(
            CharsetMatch(
                sequences,
                encoding_iana,
                mean_mess_ratio,
                bom_or_sig_available,
                cd_ratios_merged,
                decoded_payload,
            )
        )

        if (
            encoding_iana in [specified_encoding, "ascii", "utf_8"]
            and mean_mess_ratio < 0.1
        ):
            logger.debug(
                "Encoding detection: %s is most likely the one.", encoding_iana
            )
            if explain:
                logger.removeHandler(explain_handler)
                logger.setLevel(previous_logger_level)
            return CharsetMatches([results[encoding_iana]])

        if encoding_iana == sig_encoding:
            logger.debug(
                "Encoding detection: %s is most likely the one as we detected a BOM or SIG within "
                "the beginning of the sequence.",
                encoding_iana,
            )
            if explain:
                logger.removeHandler(explain_handler)
                logger.setLevel(previous_logger_level)
            return CharsetMatches([results[encoding_iana]])

    if len(results) == 0:
        if fallback_u8 or fallback_ascii or fallback_specified:
            logger.log(
                TRACE,
                "Nothing got out of the detection process. Using ASCII/UTF-8/Specified fallback.",
            )

        if fallback_specified:
            logger.debug(
                "Encoding detection: %s will be used as a fallback match",
                fallback_specified.encoding,
            )
            results.append(fallback_specified)
        elif (
            (fallback_u8 and fallback_ascii is None)
            or (
                fallback_u8
                and fallback_ascii
                and fallback_u8.fingerprint != fallback_ascii.fingerprint
            )
            or (fallback_u8 is not None)
        ):
            logger.debug("Encoding detection: utf_8 will be used as a fallback match")
            results.append(fallback_u8)
        elif fallback_ascii:
            logger.debug("Encoding detection: ascii will be used as a fallback match")
            results.append(fallback_ascii)

    if results:
        logger.debug(
            "Encoding detection: Found %s as plausible (best-candidate) for content. With %i alternatives.",
            results.best().encoding,  # type: ignore
            len(results) - 1,
        )
    else:
        logger.debug("Encoding detection: Unable to determine any suitable charset.")

    if explain:
        logger.removeHandler(explain_handler)
        logger.setLevel(previous_logger_level)

    return results


def from_fp(
    fp: BinaryIO,
    steps: int = 5,
    chunk_size: int = 512,
    threshold: float = 0.20,
    cp_isolation: List[str] = None,
    cp_exclusion: List[str] = None,
    preemptive_behaviour: bool = True,
    explain: bool = False,
) -> CharsetMatches:
    """
    Same thing than the function from_bytes but using a file pointer that is already ready.
    Will not close the file pointer.
    """
    return from_bytes(
        fp.read(),
        steps,
        chunk_size,
        threshold,
        cp_isolation,
        cp_exclusion,
        preemptive_behaviour,
        explain,
    )


def from_path(
    path: PathLike,
    steps: int = 5,
    chunk_size: int = 512,
    threshold: float = 0.20,
    cp_isolation: List[str] = None,
    cp_exclusion: List[str] = None,
    preemptive_behaviour: bool = True,
    explain: bool = False,
) -> CharsetMatches:
    """
    Same thing than the function from_bytes but with one extra step. Opening and reading given file path in binary mode.
    Can raise IOError.
    """
    with open(path, "rb") as fp:
        return from_fp(
            fp,
            steps,
            chunk_size,
            threshold,
            cp_isolation,
            cp_exclusion,
            preemptive_behaviour,
            explain,
        )


def normalize(
    path: PathLike,
    steps: int = 5,
    chunk_size: int = 512,
    threshold: float = 0.20,
    cp_isolation: List[str] = None,
    cp_exclusion: List[str] = None,
    preemptive_behaviour: bool = True,
) -> CharsetMatch:
    """
    Take a (text-based) file path and try to create another file next to it, this time using UTF-8.
    """
    results = from_path(
        path,
        steps,
        chunk_size,
        threshold,
        cp_isolation,
        cp_exclusion,
        preemptive_behaviour,
    )

    filename = basename(path)
    target_extensions = list(splitext(filename))

    if len(results) == 0:
        raise IOError(
            'Unable to normalize "{}", no encoding charset seems to fit.'.format(
                filename
            )
        )

    result = results.best()

    target_extensions[0] += "-" + result.encoding  # type: ignore

    with open(
        "{}".format(str(path).replace(filename, "".join(target_extensions))), "wb"
    ) as fp:
        fp.write(result.output())  # type: ignore

    return result  # type: ignore