File size: 24,673 Bytes
b7731cd
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
# Copyright 1999 by Jeffrey Chang.  All rights reserved.
# Copyright 2009-2018 by Peter Cock. All rights reserved.
#
# This file is part of the Biopython distribution and governed by your
# choice of the "Biopython License Agreement" or the "BSD 3-Clause License".
# Please see the LICENSE file that should have been included as part of this
# package.
"""Code for more fancy file handles.

Bio.File defines private classes used in Bio.SeqIO and Bio.SearchIO for
indexing files. These are not intended for direct use.
"""

import os
import contextlib
import itertools
import collections.abc

from abc import ABC, abstractmethod

try:
    import sqlite3
except ImportError:
    # May be missing if Python was compiled from source without its dependencies
    sqlite3 = None


@contextlib.contextmanager
def as_handle(handleish, mode="r", **kwargs):
    r"""Context manager to ensure we are using a handle.

    Context manager for arguments that can be passed to SeqIO and AlignIO read, write,
    and parse methods: either file objects or path-like objects (strings, pathlib.Path
    instances, or more generally, anything that can be handled by the builtin 'open'
    function).

    When given a path-like object, returns an open file handle to that path, with provided
    mode, which will be closed when the manager exits.

    All other inputs are returned, and are *not* closed.

    Arguments:
     - handleish  - Either a file handle or path-like object (anything which can be
                    passed to the builtin 'open' function, such as str, bytes,
                    pathlib.Path, and os.DirEntry objects)
     - mode       - Mode to open handleish (used only if handleish is a string)
     - kwargs     - Further arguments to pass to open(...)

    Examples
    --------
    >>> from Bio import File
    >>> import os
    >>> with File.as_handle('seqs.fasta', 'w') as fp:
    ...     fp.write('>test\nACGT')
    ...
    10
    >>> fp.closed
    True

    >>> handle = open('seqs.fasta', 'w')
    >>> with File.as_handle(handle) as fp:
    ...     fp.write('>test\nACGT')
    ...
    10
    >>> fp.closed
    False
    >>> fp.close()
    >>> os.remove("seqs.fasta")  # tidy up

    """
    try:
        with open(handleish, mode, **kwargs) as fp:
            yield fp
    except TypeError:
        yield handleish


def _open_for_random_access(filename):
    """Open a file in binary mode, spot if it is BGZF format etc (PRIVATE).

    This functionality is used by the Bio.SeqIO and Bio.SearchIO index
    and index_db functions.

    If the file is gzipped but not BGZF, a specific ValueError is raised.
    """
    handle = open(filename, "rb")
    magic = handle.read(2)
    handle.seek(0)

    if magic == b"\x1f\x8b":
        # This is a gzipped file, but is it BGZF?
        from . import bgzf

        try:
            # If it is BGZF, we support that
            return bgzf.BgzfReader(mode="rb", fileobj=handle)
        except ValueError as e:
            assert "BGZF" in str(e)
            # Not a BGZF file after all,
            handle.close()
            raise ValueError(
                "Gzipped files are not suitable for indexing, "
                "please use BGZF (blocked gzip format) instead."
            ) from None

    return handle


# The rest of this file defines code used in Bio.SeqIO and Bio.SearchIO
# for indexing


class _IndexedSeqFileProxy(ABC):
    """Abstract base class for file format specific random access (PRIVATE).

    This is subclasses in both Bio.SeqIO for indexing as SeqRecord
    objects, and in Bio.SearchIO for indexing QueryResult objects.

    Subclasses for each file format should define '__iter__', 'get'
    and optionally 'get_raw' methods.
    """

    @abstractmethod
    def __iter__(self):
        """Return (identifier, offset, length in bytes) tuples.

        The length can be zero where it is not implemented or not
        possible for a particular file format.
        """
        raise NotImplementedError

    @abstractmethod
    def get(self, offset):
        """Return parsed object for this entry."""
        # Most file formats with self contained records can be handled by
        # parsing StringIO(self.get_raw(offset).decode())
        raise NotImplementedError

    def get_raw(self, offset):
        """Return the raw record from the file as a bytes string (if implemented).

        If the key is not found, a KeyError exception is raised.

        This may not have been implemented for all file formats.
        """
        # Should be done by each sub-class (if possible)
        raise NotImplementedError("Not available for this file format.")


class _IndexedSeqFileDict(collections.abc.Mapping):
    """Read only dictionary interface to a sequential record file.

    This code is used in both Bio.SeqIO for indexing as SeqRecord
    objects, and in Bio.SearchIO for indexing QueryResult objects.

    Keeps the keys and associated file offsets in memory, reads the file
    to access entries as objects parsing them on demand. This approach
    is memory limited, but will work even with millions of records.

    Note duplicate keys are not allowed. If this happens, a ValueError
    exception is raised.

    As used in Bio.SeqIO, by default the SeqRecord's id string is used
    as the dictionary key. In Bio.SearchIO, the query's id string is
    used. This can be changed by supplying an optional key_function,
    a callback function which will be given the record id and must
    return the desired key. For example, this allows you to parse
    NCBI style FASTA identifiers, and extract the GI number to use
    as the dictionary key.

    Note that this dictionary is essentially read only. You cannot
    add or change values, pop values, nor clear the dictionary.
    """

    def __init__(self, random_access_proxy, key_function, repr, obj_repr):
        """Initialize the class."""
        # Use key_function=None for default value
        self._proxy = random_access_proxy
        self._key_function = key_function
        self._repr = repr
        self._obj_repr = obj_repr
        self._cached_prev_record = (None, None)  # (key, record)
        if key_function:
            offset_iter = (
                (key_function(key), offset, length)
                for (key, offset, length) in random_access_proxy
            )
        else:
            offset_iter = random_access_proxy
        offsets = {}
        for key, offset, length in offset_iter:
            # Note - we don't store the length because I want to minimise the
            # memory requirements. With the SQLite backend the length is kept
            # and is used to speed up the get_raw method (by about 3 times).
            # The length should be provided by all the current backends except
            # SFF where there is an existing Roche index we can reuse (very fast
            # but lacks the record lengths)
            # assert length or format in ["sff", "sff-trim"], \
            #       "%s at offset %i given length %r (%s format %s)" \
            #       % (key, offset, length, filename, format)
            if key in offsets:
                self._proxy._handle.close()
                raise ValueError(f"Duplicate key '{key}'")
            else:
                offsets[key] = offset
        self._offsets = offsets

    def __repr__(self):
        """Return a string representation of the File object."""
        return self._repr

    def __str__(self):
        """Create a string representation of the File object."""
        # TODO - How best to handle the __str__ for SeqIO and SearchIO?
        if self:
            return f"{{{list(self.keys())[0]!r} : {self._obj_repr}(...), ...}}"
        else:
            return "{}"

    def __len__(self):
        """Return the number of records."""
        return len(self._offsets)

    def __iter__(self):
        """Iterate over the keys."""
        return iter(self._offsets)

    def __getitem__(self, key):
        """Return record for the specified key.

        As an optimization when repeatedly asked to look up the same record,
        the key and record are cached so that if the *same* record is
        requested next time, it can be returned without going to disk.
        """
        if key == self._cached_prev_record[0]:
            return self._cached_prev_record[1]
        # Pass the offset to the proxy
        record = self._proxy.get(self._offsets[key])
        if self._key_function:
            key2 = self._key_function(record.id)
        else:
            key2 = record.id
        if key != key2:
            raise ValueError(f"Key did not match ({key} vs {key2})")
        self._cached_prev_record = (key, record)
        return record

    def get_raw(self, key):
        """Return the raw record from the file as a bytes string.

        If the key is not found, a KeyError exception is raised.
        """
        # Pass the offset to the proxy
        return self._proxy.get_raw(self._offsets[key])

    def close(self):
        """Close the file handle being used to read the data.

        Once called, further use of the index won't work. The sole purpose
        of this method is to allow explicit handle closure - for example
        if you wish to delete the file, on Windows you must first close
        all open handles to that file.
        """
        self._proxy._handle.close()


class _SQLiteManySeqFilesDict(_IndexedSeqFileDict):
    """Read only dictionary interface to many sequential record files.

    This code is used in both Bio.SeqIO for indexing as SeqRecord
    objects, and in Bio.SearchIO for indexing QueryResult objects.

    Keeps the keys, file-numbers and offsets in an SQLite database. To access
    a record by key, reads from the offset in the appropriate file and then
    parses the record into an object.

    There are OS limits on the number of files that can be open at once,
    so a pool are kept. If a record is required from a closed file, then
    one of the open handles is closed first.
    """

    def __init__(
        self,
        index_filename,
        filenames,
        proxy_factory,
        fmt,
        key_function,
        repr,
        max_open=10,
    ):
        """Initialize the class."""
        # TODO? - Don't keep filename list in memory (just in DB)?
        # Should save a chunk of memory if dealing with 1000s of files.
        # Furthermore could compare a generator to the DB on reloading
        # (no need to turn it into a list)

        if sqlite3 is None:
            # Python was compiled without sqlite3 support
            from Bio import MissingPythonDependencyError

            raise MissingPythonDependencyError(
                "Python was compiled without the sqlite3 module"
            )
        if filenames is not None:
            filenames = list(filenames)  # In case it was a generator

        # Cache the arguments as private variables
        self._index_filename = index_filename
        self._filenames = filenames
        self._format = fmt
        self._key_function = key_function
        self._proxy_factory = proxy_factory
        self._repr = repr
        self._max_open = max_open
        self._proxies = {}

        # Note if using SQLite :memory: trick index filename, this will
        # give $PWD as the relative path (which is fine).
        self._relative_path = os.path.abspath(os.path.dirname(index_filename))

        if os.path.isfile(index_filename):
            self._load_index()
        else:
            self._build_index()

    def _load_index(self):
        """Call from __init__ to re-use an existing index (PRIVATE)."""
        index_filename = self._index_filename
        relative_path = self._relative_path
        filenames = self._filenames
        fmt = self._format
        proxy_factory = self._proxy_factory

        con = sqlite3.dbapi2.connect(index_filename, check_same_thread=False)
        self._con = con
        # Check the count...
        try:
            (count,) = con.execute(
                "SELECT value FROM meta_data WHERE key=?;", ("count",)
            ).fetchone()
            self._length = int(count)
            if self._length == -1:
                con.close()
                raise ValueError("Unfinished/partial database") from None

            # use MAX(_ROWID_) to obtain the number of sequences in the database
            # using COUNT(key) is quite slow in SQLITE
            # (https://stackoverflow.com/questions/8988915/sqlite-count-slow-on-big-tables)
            (count,) = con.execute("SELECT MAX(_ROWID_) FROM offset_data;").fetchone()
            if self._length != int(count):
                con.close()
                raise ValueError(
                    "Corrupt database? %i entries not %i" % (int(count), self._length)
                ) from None
            (self._format,) = con.execute(
                "SELECT value FROM meta_data WHERE key=?;", ("format",)
            ).fetchone()
            if fmt and fmt != self._format:
                con.close()
                raise ValueError(
                    f"Index file says format {self._format}, not {fmt}"
                ) from None
            try:
                (filenames_relative_to_index,) = con.execute(
                    "SELECT value FROM meta_data WHERE key=?;",
                    ("filenames_relative_to_index",),
                ).fetchone()
                filenames_relative_to_index = (
                    filenames_relative_to_index.upper() == "TRUE"
                )
            except TypeError:
                # Original behaviour, assume if meta_data missing
                filenames_relative_to_index = False
            self._filenames = [
                row[0]
                for row in con.execute(
                    "SELECT name FROM file_data ORDER BY file_number;"
                ).fetchall()
            ]
            if filenames_relative_to_index:
                # Not implicitly relative to $PWD, explicitly relative to index file
                relative_path = os.path.abspath(os.path.dirname(index_filename))
                tmp = []
                for f in self._filenames:
                    if os.path.isabs(f):
                        tmp.append(f)
                    else:
                        # Would be stored with Unix / path separator, so convert
                        # it to the local OS path separator here:
                        tmp.append(
                            os.path.join(relative_path, f.replace("/", os.path.sep))
                        )
                self._filenames = tmp
                del tmp
            if filenames and len(filenames) != len(self._filenames):
                con.close()
                raise ValueError(
                    "Index file says %i files, not %i"
                    % (len(self._filenames), len(filenames))
                ) from None
            if filenames and filenames != self._filenames:
                for old, new in zip(self._filenames, filenames):
                    # Want exact match (after making relative to the index above)
                    if os.path.abspath(old) != os.path.abspath(new):
                        con.close()
                        if filenames_relative_to_index:
                            raise ValueError(
                                "Index file has different filenames, e.g. %r != %r"
                                % (os.path.abspath(old), os.path.abspath(new))
                            ) from None
                        else:
                            raise ValueError(
                                "Index file has different filenames "
                                "[This is an old index where any relative paths "
                                "were relative to the original working directory]. "
                                "e.g. %r != %r"
                                % (os.path.abspath(old), os.path.abspath(new))
                            ) from None
                # Filenames are equal (after imposing abspath)
        except sqlite3.OperationalError as err:
            con.close()
            raise ValueError(f"Not a Biopython index database? {err}") from None
        # Now we have the format (from the DB if not given to us),
        if not proxy_factory(self._format):
            con.close()
            raise ValueError(f"Unsupported format '{self._format}'")

    def _build_index(self):
        """Call from __init__ to create a new index (PRIVATE)."""
        index_filename = self._index_filename
        relative_path = self._relative_path
        filenames = self._filenames
        fmt = self._format
        key_function = self._key_function
        proxy_factory = self._proxy_factory
        max_open = self._max_open
        random_access_proxies = self._proxies

        if not fmt or not filenames:
            raise ValueError(
                f"Filenames to index and format required to build {index_filename!r}"
            )
        if not proxy_factory(fmt):
            raise ValueError(f"Unsupported format '{fmt}'")
        # Create the index
        con = sqlite3.dbapi2.connect(index_filename)
        self._con = con
        # print("Creating index")
        # Sqlite PRAGMA settings for speed
        con.execute("PRAGMA synchronous=OFF")
        con.execute("PRAGMA locking_mode=EXCLUSIVE")
        # Don't index the key column until the end (faster)
        # con.execute("CREATE TABLE offset_data (key TEXT PRIMARY KEY, "
        #             "offset INTEGER);")
        con.execute("CREATE TABLE meta_data (key TEXT, value TEXT);")
        con.execute("INSERT INTO meta_data (key, value) VALUES (?,?);", ("count", -1))
        con.execute("INSERT INTO meta_data (key, value) VALUES (?,?);", ("format", fmt))
        con.execute(
            "INSERT INTO meta_data (key, value) VALUES (?,?);",
            ("filenames_relative_to_index", "True"),
        )
        # TODO - Record the file size and modified date?
        con.execute("CREATE TABLE file_data (file_number INTEGER, name TEXT);")
        con.execute(
            "CREATE TABLE offset_data (key TEXT, "
            "file_number INTEGER, offset INTEGER, length INTEGER);"
        )
        count = 0
        for file_index, filename in enumerate(filenames):
            # Default to storing as an absolute path,
            f = os.path.abspath(filename)
            if not os.path.isabs(filename) and not os.path.isabs(index_filename):
                # Since user gave BOTH filename & index as relative paths,
                # we will store this relative to the index file even though
                # if it may now start ../ (meaning up a level)
                # Note for cross platform use (e.g. shared drive over SAMBA),
                # convert any Windows slash into Unix style for rel paths.
                f = os.path.relpath(filename, relative_path).replace(os.path.sep, "/")
            elif (os.path.dirname(os.path.abspath(filename)) + os.path.sep).startswith(
                relative_path + os.path.sep
            ):
                # Since sequence file is in same directory or sub directory,
                # might as well make this into a relative path:
                f = os.path.relpath(filename, relative_path).replace(os.path.sep, "/")
                assert not f.startswith("../"), f
            # print("DEBUG - storing %r as [%r] %r" % (filename, relative_path, f))
            con.execute(
                "INSERT INTO file_data (file_number, name) VALUES (?,?);",
                (file_index, f),
            )
            random_access_proxy = proxy_factory(fmt, filename)
            if key_function:
                offset_iter = (
                    (key_function(key), file_index, offset, length)
                    for (key, offset, length) in random_access_proxy
                )
            else:
                offset_iter = (
                    (key, file_index, offset, length)
                    for (key, offset, length) in random_access_proxy
                )
            while True:
                batch = list(itertools.islice(offset_iter, 100))
                if not batch:
                    break
                # print("Inserting batch of %i offsets, %s ... %s"
                #       % (len(batch), batch[0][0], batch[-1][0]))
                con.executemany(
                    "INSERT INTO offset_data (key,file_number,offset,length) VALUES (?,?,?,?);",
                    batch,
                )
                con.commit()
                count += len(batch)
            if len(random_access_proxies) < max_open:
                random_access_proxies[file_index] = random_access_proxy
            else:
                random_access_proxy._handle.close()
        self._length = count
        # print("About to index %i entries" % count)
        try:
            con.execute(
                "CREATE UNIQUE INDEX IF NOT EXISTS key_index ON offset_data(key);"
            )
        except sqlite3.IntegrityError as err:
            self._proxies = random_access_proxies
            self.close()
            con.close()
            raise ValueError(f"Duplicate key? {err}") from None
        con.execute("PRAGMA locking_mode=NORMAL")
        con.execute("UPDATE meta_data SET value = ? WHERE key = ?;", (count, "count"))
        con.commit()
        # print("Index created")

    def __repr__(self):
        return self._repr

    def __contains__(self, key):
        return bool(
            self._con.execute(
                "SELECT key FROM offset_data WHERE key=?;", (key,)
            ).fetchone()
        )

    def __len__(self):
        """Return the number of records indexed."""
        return self._length
        # return self._con.execute("SELECT COUNT(key) FROM offset_data;").fetchone()[0]

    def __iter__(self):
        """Iterate over the keys."""
        for row in self._con.execute(
            "SELECT key FROM offset_data ORDER BY file_number, offset;"
        ):
            yield str(row[0])

    def __getitem__(self, key):
        """Return record for the specified key."""
        # Pass the offset to the proxy
        row = self._con.execute(
            "SELECT file_number, offset FROM offset_data WHERE key=?;", (key,)
        ).fetchone()
        if not row:
            raise KeyError
        file_number, offset = row
        proxies = self._proxies
        if file_number in proxies:
            record = proxies[file_number].get(offset)
        else:
            if len(proxies) >= self._max_open:
                # Close an old handle...
                proxies.popitem()[1]._handle.close()
            # Open a new handle...
            proxy = self._proxy_factory(self._format, self._filenames[file_number])
            record = proxy.get(offset)
            proxies[file_number] = proxy
        if self._key_function:
            key2 = self._key_function(record.id)
        else:
            key2 = record.id
        if key != key2:
            raise ValueError(f"Key did not match ({key} vs {key2})")
        return record

    def get_raw(self, key):
        """Return the raw record from the file as a bytes string.

        If the key is not found, a KeyError exception is raised.
        """
        # Pass the offset to the proxy
        row = self._con.execute(
            "SELECT file_number, offset, length FROM offset_data WHERE key=?;", (key,)
        ).fetchone()
        if not row:
            raise KeyError
        file_number, offset, length = row
        proxies = self._proxies
        if file_number in proxies:
            if length:
                # Shortcut if we have the length
                h = proxies[file_number]._handle
                h.seek(offset)
                return h.read(length)
            else:
                return proxies[file_number].get_raw(offset)
        else:
            # This code is duplicated from __getitem__ to avoid a function call
            if len(proxies) >= self._max_open:
                # Close an old handle...
                proxies.popitem()[1]._handle.close()
            # Open a new handle...
            proxy = self._proxy_factory(self._format, self._filenames[file_number])
            proxies[file_number] = proxy
            if length:
                # Shortcut if we have the length
                h = proxy._handle
                h.seek(offset)
                return h.read(length)
            else:
                return proxy.get_raw(offset)

    def close(self):
        """Close any open file handles."""
        proxies = self._proxies
        while proxies:
            proxies.popitem()[1]._handle.close()