summaryrefslogtreecommitdiff
path: root/paramiko/sftp_file.py
blob: 03d67b332db40a381c6e674c0f2d35950bf4a667 (plain)
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
# Copyright (C) 2003-2007  Robey Pointer <robeypointer@gmail.com>
#
# This file is part of paramiko.
#
# Paramiko is free software; you can redistribute it and/or modify it under the
# terms of the GNU Lesser General Public License as published by the Free
# Software Foundation; either version 2.1 of the License, or (at your option)
# any later version.
#
# Paramiko is distributed in the hope that it will be useful, but WITHOUT ANY
# WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
# A PARTICULAR PURPOSE.  See the GNU Lesser General Public License for more
# details.
#
# You should have received a copy of the GNU Lesser General Public License
# along with Paramiko; if not, write to the Free Software Foundation, Inc.,
# 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA.

"""
SFTP file object
"""

from __future__ import with_statement

from binascii import hexlify
from collections import deque
import socket
import threading
import time
from paramiko.common import DEBUG

from paramiko.file import BufferedFile
from paramiko.py3compat import long
from paramiko.sftp import CMD_CLOSE, CMD_READ, CMD_DATA, SFTPError, CMD_WRITE, \
    CMD_STATUS, CMD_FSTAT, CMD_ATTRS, CMD_FSETSTAT, CMD_EXTENDED
from paramiko.sftp_attr import SFTPAttributes


class SFTPFile (BufferedFile):
    """
    Proxy object for a file on the remote server, in client mode SFTP.

    Instances of this class may be used as context managers in the same way
    that built-in Python file objects are.
    """

    # Some sftp servers will choke if you send read/write requests larger than
    # this size.
    MAX_REQUEST_SIZE = 32768

    def __init__(self, sftp, handle, mode='r', bufsize=-1):
        BufferedFile.__init__(self)
        self.sftp = sftp
        self.handle = handle
        BufferedFile._set_mode(self, mode, bufsize)
        self.pipelined = False
        self._prefetching = False
        self._prefetch_done = False
        self._prefetch_data = {}
        self._prefetch_extents = {}
        self._prefetch_lock = threading.Lock()
        self._saved_exception = None
        self._reqs = deque()

    def __del__(self):
        self._close(async=True)
    
    def close(self):
        """
        Close the file.
        """
        self._close(async=False)
        
    def _close(self, async=False):
        # We allow double-close without signaling an error, because real
        # Python file objects do.  However, we must protect against actually
        # sending multiple CMD_CLOSE packets, because after we close our
        # handle, the same handle may be re-allocated by the server, and we
        # may end up mysteriously closing some random other file.  (This is
        # especially important because we unconditionally call close() from
        # __del__.)
        if self._closed:
            return
        self.sftp._log(DEBUG, 'close(%s)' % hexlify(self.handle))
        if self.pipelined:
            self.sftp._finish_responses(self)
        BufferedFile.close(self)
        try:
            if async:
                # GC'd file handle could be called from an arbitrary thread -- don't wait for a response
                self.sftp._async_request(type(None), CMD_CLOSE, self.handle)
            else:
                self.sftp._request(CMD_CLOSE, self.handle)
        except EOFError:
            # may have outlived the Transport connection
            pass
        except (IOError, socket.error):
            # may have outlived the Transport connection
            pass

    def _data_in_prefetch_requests(self, offset, size):
        k = [x for x in list(self._prefetch_extents.values()) if x[0] <= offset]
        if len(k) == 0:
            return False
        k.sort(key=lambda x: x[0])
        buf_offset, buf_size = k[-1]
        if buf_offset + buf_size <= offset:
            # prefetch request ends before this one begins
            return False
        if buf_offset + buf_size >= offset + size:
            # inclusive
            return True
        # well, we have part of the request.  see if another chunk has the rest.
        return self._data_in_prefetch_requests(buf_offset + buf_size, offset + size - buf_offset - buf_size)
    
    def _data_in_prefetch_buffers(self, offset):
        """
        if a block of data is present in the prefetch buffers, at the given
        offset, return the offset of the relevant prefetch buffer.  otherwise,
        return None.  this guarantees nothing about the number of bytes
        collected in the prefetch buffer so far.
        """
        k = [i for i in self._prefetch_data.keys() if i <= offset]
        if len(k) == 0:
            return None
        index = max(k)
        buf_offset = offset - index
        if buf_offset >= len(self._prefetch_data[index]):
            # it's not here
            return None
        return index
        
    def _read_prefetch(self, size):
        """
        read data out of the prefetch buffer, if possible.  if the data isn't
        in the buffer, return None.  otherwise, behaves like a normal read.
        """
        # while not closed, and haven't fetched past the current position, and haven't reached EOF...
        while True:
            offset = self._data_in_prefetch_buffers(self._realpos)
            if offset is not None:
                break
            if self._prefetch_done or self._closed:
                break
            self.sftp._read_response()
            self._check_exception()
        if offset is None:
            self._prefetching = False
            return None
        prefetch = self._prefetch_data[offset]
        del self._prefetch_data[offset]
        
        buf_offset = self._realpos - offset
        if buf_offset > 0:
            self._prefetch_data[offset] = prefetch[:buf_offset]
            prefetch = prefetch[buf_offset:]
        if size < len(prefetch):
            self._prefetch_data[self._realpos + size] = prefetch[size:]
            prefetch = prefetch[:size]
        return prefetch
        
    def _read(self, size):
        size = min(size, self.MAX_REQUEST_SIZE)
        if self._prefetching:
            data = self._read_prefetch(size)
            if data is not None:
                return data
        t, msg = self.sftp._request(CMD_READ, self.handle, long(self._realpos), int(size))
        if t != CMD_DATA:
            raise SFTPError('Expected data')
        return msg.get_string()

    def _write(self, data):
        # may write less than requested if it would exceed max packet size
        chunk = min(len(data), self.MAX_REQUEST_SIZE)
        self._reqs.append(self.sftp._async_request(type(None), CMD_WRITE, self.handle, long(self._realpos), data[:chunk]))
        if not self.pipelined or (len(self._reqs) > 100 and self.sftp.sock.recv_ready()):
            while len(self._reqs):
                req = self._reqs.popleft()
                t, msg = self.sftp._read_response(req)
                if t != CMD_STATUS:
                    raise SFTPError('Expected status')
                # convert_status already called
        return chunk

    def settimeout(self, timeout):
        """
        Set a timeout on read/write operations on the underlying socket or
        ssh `.Channel`.

        :param float timeout:
            seconds to wait for a pending read/write operation before raising
            ``socket.timeout``, or ``None`` for no timeout

        .. seealso:: `.Channel.settimeout`
        """
        self.sftp.sock.settimeout(timeout)

    def gettimeout(self):
        """
        Returns the timeout in seconds (as a `float`) associated with the
        socket or ssh `.Channel` used for this file.

        .. seealso:: `.Channel.gettimeout`
        """
        return self.sftp.sock.gettimeout()

    def setblocking(self, blocking):
        """
        Set blocking or non-blocking mode on the underiying socket or ssh
        `.Channel`.

        :param int blocking:
            0 to set non-blocking mode; non-0 to set blocking mode.

        .. seealso:: `.Channel.setblocking`
        """
        self.sftp.sock.setblocking(blocking)

    def seek(self, offset, whence=0):
        self.flush()
        if whence == self.SEEK_SET:
            self._realpos = self._pos = offset
        elif whence == self.SEEK_CUR:
            self._pos += offset
            self._realpos = self._pos
        else:
            self._realpos = self._pos = self._get_size() + offset
        self._rbuffer = bytes()

    def stat(self):
        """
        Retrieve information about this file from the remote system.  This is
        exactly like `.SFTPClient.stat`, except that it operates on an
        already-open file.

        :return: an `.SFTPAttributes` object containing attributes about this file.
        """
        t, msg = self.sftp._request(CMD_FSTAT, self.handle)
        if t != CMD_ATTRS:
            raise SFTPError('Expected attributes')
        return SFTPAttributes._from_msg(msg)

    def chmod(self, mode):
        """
        Change the mode (permissions) of this file.  The permissions are
        unix-style and identical to those used by Python's `os.chmod`
        function.

        :param int mode: new permissions
        """
        self.sftp._log(DEBUG, 'chmod(%s, %r)' % (hexlify(self.handle), mode))
        attr = SFTPAttributes()
        attr.st_mode = mode
        self.sftp._request(CMD_FSETSTAT, self.handle, attr)
        
    def chown(self, uid, gid):
        """
        Change the owner (``uid``) and group (``gid``) of this file.  As with
        Python's `os.chown` function, you must pass both arguments, so if you
        only want to change one, use `stat` first to retrieve the current
        owner and group.

        :param int uid: new owner's uid
        :param int gid: new group id
        """
        self.sftp._log(DEBUG, 'chown(%s, %r, %r)' % (hexlify(self.handle), uid, gid))
        attr = SFTPAttributes()
        attr.st_uid, attr.st_gid = uid, gid
        self.sftp._request(CMD_FSETSTAT, self.handle, attr)

    def utime(self, times):
        """
        Set the access and modified times of this file.  If
        ``times`` is ``None``, then the file's access and modified times are set
        to the current time.  Otherwise, ``times`` must be a 2-tuple of numbers,
        of the form ``(atime, mtime)``, which is used to set the access and
        modified times, respectively.  This bizarre API is mimicked from Python
        for the sake of consistency -- I apologize.

        :param tuple times:
            ``None`` or a tuple of (access time, modified time) in standard
            internet epoch time (seconds since 01 January 1970 GMT)
        """
        if times is None:
            times = (time.time(), time.time())
        self.sftp._log(DEBUG, 'utime(%s, %r)' % (hexlify(self.handle), times))
        attr = SFTPAttributes()
        attr.st_atime, attr.st_mtime = times
        self.sftp._request(CMD_FSETSTAT, self.handle, attr)

    def truncate(self, size):
        """
        Change the size of this file.  This usually extends
        or shrinks the size of the file, just like the ``truncate()`` method on
        Python file objects.
        
        :param size: the new size of the file
        :type size: int or long
        """
        self.sftp._log(DEBUG, 'truncate(%s, %r)' % (hexlify(self.handle), size))
        attr = SFTPAttributes()
        attr.st_size = size
        self.sftp._request(CMD_FSETSTAT, self.handle, attr)
    
    def check(self, hash_algorithm, offset=0, length=0, block_size=0):
        """
        Ask the server for a hash of a section of this file.  This can be used
        to verify a successful upload or download, or for various rsync-like
        operations.
        
        The file is hashed from ``offset``, for ``length`` bytes.  If ``length``
        is 0, the remainder of the file is hashed.  Thus, if both ``offset``
        and ``length`` are zero, the entire file is hashed.
        
        Normally, ``block_size`` will be 0 (the default), and this method will
        return a byte string representing the requested hash (for example, a
        string of length 16 for MD5, or 20 for SHA-1).  If a non-zero
        ``block_size`` is given, each chunk of the file (from ``offset`` to
        ``offset + length``) of ``block_size`` bytes is computed as a separate
        hash.  The hash results are all concatenated and returned as a single
        string.
        
        For example, ``check('sha1', 0, 1024, 512)`` will return a string of
        length 40.  The first 20 bytes will be the SHA-1 of the first 512 bytes
        of the file, and the last 20 bytes will be the SHA-1 of the next 512
        bytes.
        
        :param str hash_algorithm:
            the name of the hash algorithm to use (normally ``"sha1"`` or
            ``"md5"``)
        :param offset:
            offset into the file to begin hashing (0 means to start from the
            beginning)
        :type offset: int or long
        :param length:
            number of bytes to hash (0 means continue to the end of the file)
        :type length: int or long
        :param int block_size:
            number of bytes to hash per result (must not be less than 256; 0
            means to compute only one hash of the entire segment)
        :type block_size: int
        :return:
            `str` of bytes representing the hash of each block, concatenated
            together
        
        :raises IOError: if the server doesn't support the "check-file"
            extension, or possibly doesn't support the hash algorithm
            requested
            
        .. note:: Many (most?) servers don't support this extension yet.
        
        .. versionadded:: 1.4
        """
        t, msg = self.sftp._request(CMD_EXTENDED, 'check-file', self.handle,
                                    hash_algorithm, long(offset), long(length), block_size)
        ext = msg.get_text()
        alg = msg.get_text()
        data = msg.get_remainder()
        return data
    
    def set_pipelined(self, pipelined=True):
        """
        Turn on/off the pipelining of write operations to this file.  When
        pipelining is on, paramiko won't wait for the server response after
        each write operation.  Instead, they're collected as they come in. At
        the first non-write operation (including `.close`), all remaining
        server responses are collected.  This means that if there was an error
        with one of your later writes, an exception might be thrown from within
        `.close` instead of `.write`.
        
        By default, files are not pipelined.
        
        :param bool pipelined:
            ``True`` if pipelining should be turned on for this file; ``False``
            otherwise
        
        .. versionadded:: 1.5
        """
        self.pipelined = pipelined
    
    def prefetch(self):
        """
        Pre-fetch the remaining contents of this file in anticipation of future
        `.read` calls.  If reading the entire file, pre-fetching can
        dramatically improve the download speed by avoiding roundtrip latency.
        The file's contents are incrementally buffered in a background thread.
        
        The prefetched data is stored in a buffer until read via the `.read`
        method.  Once data has been read, it's removed from the buffer.  The
        data may be read in a random order (using `.seek`); chunks of the
        buffer that haven't been read will continue to be buffered.

        .. versionadded:: 1.5.1
        """
        size = self.stat().st_size
        # queue up async reads for the rest of the file
        chunks = []
        n = self._realpos
        while n < size:
            chunk = min(self.MAX_REQUEST_SIZE, size - n)
            chunks.append((n, chunk))
            n += chunk
        if len(chunks) > 0:
            self._start_prefetch(chunks)
    
    def readv(self, chunks):
        """
        Read a set of blocks from the file by (offset, length).  This is more
        efficient than doing a series of `.seek` and `.read` calls, since the
        prefetch machinery is used to retrieve all the requested blocks at
        once.
        
        :param chunks:
            a list of (offset, length) tuples indicating which sections of the
            file to read
        :type chunks: list(tuple(long, int))
        :return: a list of blocks read, in the same order as in ``chunks``
        
        .. versionadded:: 1.5.4
        """
        self.sftp._log(DEBUG, 'readv(%s, %r)' % (hexlify(self.handle), chunks))

        read_chunks = []
        for offset, size in chunks:
            # don't fetch data that's already in the prefetch buffer
            if self._data_in_prefetch_buffers(offset) or self._data_in_prefetch_requests(offset, size):
                continue

            # break up anything larger than the max read size
            while size > 0:
                chunk_size = min(size, self.MAX_REQUEST_SIZE)
                read_chunks.append((offset, chunk_size))
                offset += chunk_size
                size -= chunk_size

        self._start_prefetch(read_chunks)
        # now we can just devolve to a bunch of read()s :)
        for x in chunks:
            self.seek(x[0])
            yield self.read(x[1])

    ###  internals...

    def _get_size(self):
        try:
            return self.stat().st_size
        except:
            return 0

    def _start_prefetch(self, chunks):
        self._prefetching = True
        self._prefetch_done = False

        t = threading.Thread(target=self._prefetch_thread, args=(chunks,))
        t.setDaemon(True)
        t.start()
        
    def _prefetch_thread(self, chunks):
        # do these read requests in a temporary thread because there may be
        # a lot of them, so it may block.
        for offset, length in chunks:
            with self._prefetch_lock:
                num = self.sftp._async_request(self, CMD_READ, self.handle, long(offset), int(length))
                self._prefetch_extents[num] = (offset, length)

    def _async_response(self, t, msg, num):
        if t == CMD_STATUS:
            # save exception and re-raise it on next file operation
            try:
                self.sftp._convert_status(msg)
            except Exception as e:
                self._saved_exception = e
            return
        if t != CMD_DATA:
            raise SFTPError('Expected data')
        data = msg.get_string()
        with self._prefetch_lock:
            offset, length = self._prefetch_extents[num]
            self._prefetch_data[offset] = data
            del self._prefetch_extents[num]
            if len(self._prefetch_extents) == 0:
                self._prefetch_done = True
    
    def _check_exception(self):
        """if there's a saved exception, raise & clear it"""
        if self._saved_exception is not None:
            x = self._saved_exception
            self._saved_exception = None
            raise x
            
    def __enter__(self):
        return self

    def __exit__(self, type, value, traceback):
        self.close()