v3bwfile.py 17.1 KB
Newer Older
juga's avatar
juga committed
1
2
3
4
# -*- coding: utf-8 -*-
"""Classes and functions that create the bandwidth measurements document
(v3bw) used by bandwidth authorities."""

juga's avatar
juga committed
5
import copy
juga's avatar
juga committed
6
import logging
juga's avatar
juga committed
7
import os
juga's avatar
juga committed
8
from statistics import median, mean
juga's avatar
juga committed
9

10
from sbws import __version__
juga's avatar
juga committed
11
from sbws.globals import SPEC_VERSION, BW_LINE_SIZE, SCALE_CONSTANT
juga's avatar
juga committed
12
from sbws.lib.resultdump import ResultSuccess, _ResultType
13
from sbws.util.filelock import DirectoryLock
Matt Traudt's avatar
Matt Traudt committed
14
from sbws.util.timestamp import now_isodt_str, unixts_to_isodt_str
15
from sbws.util.state import State
juga's avatar
juga committed
16
17
18

log = logging.getLogger(__name__)

19
LINE_SEP = '\n'
20
21
22
23
24
25
KEYVALUE_SEP_V110 = '='
KEYVALUE_SEP_V200 = ' '
# List of the extra KeyValues accepted by the class
EXTRA_ARG_KEYVALUES = ['software', 'software_version', 'file_created',
                       'earliest_bandwidth', 'generator_started']
# List of all unordered KeyValues currently being used to generate the file
juga's avatar
juga committed
26
UNORDERED_KEYVALUES = EXTRA_ARG_KEYVALUES + ['latest_bandwidth']
27
28
# List of all the KeyValues currently being used to generate the file
ALL_KEYVALUES = ['version'] + UNORDERED_KEYVALUES
29
TERMINATOR = '===='
juga's avatar
juga committed
30
31
# Num header lines in v1.1.0 using all the KeyValues
NUM_LINES_HEADER_V110 = len(ALL_KEYVALUES) + 2
32
33
LINE_TERMINATOR = TERMINATOR + LINE_SEP

juga's avatar
juga committed
34
35
# KeyValue separator in Bandwidth Lines
BW_KEYVALUE_SEP_V110 = ' '
juga's avatar
juga committed
36
BW_EXTRA_ARG_KEYVALUES = ['master_key_ed25519', 'nick', 'rtt', 'time',
juga's avatar
juga committed
37
                          'success', 'error_stream', 'error_circ',
juga's avatar
juga committed
38
                          'error_misc']
39
BW_KEYVALUES_INT = ['bw', 'rtt', 'success', 'error_stream',
juga's avatar
juga committed
40
                    'error_circ', 'error_misc']
juga's avatar
juga committed
41
42
43
44
45
46
47
48
49
50
51
52
BW_KEYVALUES = ['node_id', 'bw'] + BW_EXTRA_ARG_KEYVALUES


def num_results_of_type(results, type_str):
    return len([r for r in results if r.type == type_str])


# Better way to use enums?
def result_type_to_key(type_str):
    return type_str.replace('-', '_')


53
class V3BWHeader(object):
juga's avatar
juga committed
54
55
56
57
    """
    Create a bandwidth measurements (V3bw) header
    following bandwidth measurements document spec version 1.1.0.

58
    :param str timestamp: timestamp in Unix Epoch seconds of the most recent
59
        generator result.
juga's avatar
juga committed
60
61
62
    :param str version: the spec version
    :param str software: the name of the software that generates this
    :param str software_version: the version of the software
63
    :param dict kwargs: extra headers. Currently supported:
juga's avatar
juga committed
64

65
66
67
68
        - earliest_bandwidth: str, ISO 8601 timestamp in UTC time zone
          when the first bandwidth was obtained
        - generator_started: str, ISO 8601 timestamp in UTC time zone
          when the generator started
juga's avatar
juga committed
69
    """
70
    def __init__(self, timestamp, **kwargs):
juga's avatar
juga committed
71
72
73
74
        assert isinstance(timestamp, str)
        for v in kwargs.values():
            assert isinstance(v, str)
        self.timestamp = timestamp
75
76
77
78
        # KeyValues with default value when not given by kwargs
        self.version = kwargs.get('version', SPEC_VERSION)
        self.software = kwargs.get('software', 'sbws')
        self.software_version = kwargs.get('software_version', __version__)
79
        self.file_created = kwargs.get('file_created', now_isodt_str())
juga's avatar
juga committed
80
        # latest_bandwidth should not be in kwargs, since it MUST be the
81
        # same as timestamp
juga's avatar
juga committed
82
        self.latest_bandwidth = unixts_to_isodt_str(timestamp)
83
84
        [setattr(self, k, v) for k, v in kwargs.items()
         if k in EXTRA_ARG_KEYVALUES]
juga's avatar
juga committed
85

86
87
88
89
90
    def __str__(self):
        if self.version == '1.1.0':
            return self.strv110
        return self.strv200

juga's avatar
juga committed
91
    @classmethod
juga's avatar
juga committed
92
    def from_results(cls, results, state_fpath=''):
juga's avatar
juga committed
93
94
95
        kwargs = dict()
        latest_bandwidth = cls.latest_bandwidth_from_results(results)
        earliest_bandwidth = cls.earliest_bandwidth_from_results(results)
juga's avatar
juga committed
96
        generator_started = cls.generator_started_from_file(state_fpath)
juga's avatar
juga committed
97
98
99
100
101
102
103
104
        timestamp = str(latest_bandwidth)
        kwargs['latest_bandwidth'] = unixts_to_isodt_str(latest_bandwidth)
        kwargs['earliest_bandwidth'] = unixts_to_isodt_str(earliest_bandwidth)
        if generator_started is not None:
            kwargs['generator_started'] = generator_started
        h = cls(timestamp, **kwargs)
        return h

105
106
107
108
    @classmethod
    def from_lines_v110(cls, lines):
        """
        :param list lines: list of lines to parse
109
        :returns: tuple of V3BWHeader object and non-header lines
110
111
112
113
        """
        assert isinstance(lines, list)
        try:
            index_terminator = lines.index(TERMINATOR)
Matt Traudt's avatar
Matt Traudt committed
114
        except ValueError:
115
116
117
            # is not a bw file or is v100
            log.warn('Terminator is not in lines')
            return None
118
        ts = lines[0]
119
        kwargs = dict([l.split(KEYVALUE_SEP_V110)
120
                       for l in lines[:index_terminator]
121
                       if l.split(KEYVALUE_SEP_V110)[0] in ALL_KEYVALUES])
122
        h = cls(ts, **kwargs)
123
124
       # last line is new line
        return h, lines[index_terminator + 1:-1]
125
126
127
128

    @classmethod
    def from_text_v110(self, text):
        """
129
        :param str text: text to parse
130
        :returns: tuple of V3BWHeader object and non-header lines
131
132
133
        """
        assert isinstance(text, str)
        return self.from_lines_v110(text.split(LINE_SEP))
juga's avatar
juga committed
134
135

    @staticmethod
juga's avatar
juga committed
136
    def generator_started_from_file(state_fpath):
137
138
139
140
        '''
        ISO formatted timestamp for the time when the scanner process most
        recently started.
        '''
juga's avatar
juga committed
141
        state = State(state_fpath)
142
143
144
145
        if 'scanner_started' in state:
            return state['scanner_started']
        else:
            return None
juga's avatar
juga committed
146
147

    @staticmethod
juga's avatar
juga committed
148
    def latest_bandwidth_from_results(results):
juga's avatar
juga committed
149
        return round(max([r.time for fp in results for r in results[fp]]))
juga's avatar
juga committed
150
151
152

    @staticmethod
    def earliest_bandwidth_from_results(results):
juga's avatar
juga committed
153
        return round(min([r.time for fp in results for r in results[fp]]))
juga's avatar
juga committed
154

juga's avatar
juga committed
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
    @property
    def keyvalue_unordered_tuple_ls(self):
        """Return list of KeyValue tuples that do not have specific order."""
        # sort the list to generate determinist headers
        keyvalue_tuple_ls = sorted([(k, v) for k, v in self.__dict__.items()
                                    if k in UNORDERED_KEYVALUES])
        return keyvalue_tuple_ls

    @property
    def keyvalue_tuple_ls(self):
        """Return list of all KeyValue tuples"""
        return [('version', self.version)] + self.keyvalue_unordered_tuple_ls

    @property
    def keyvalue_v110str_ls(self):
        """Return KeyValue list of strings following spec v1.1.0."""
        keyvalues = [self.timestamp] + [KEYVALUE_SEP_V110.join([k, v])
                                        for k, v in self.keyvalue_tuple_ls]
        return keyvalues

    @property
    def strv110(self):
        """Return header string following spec v1.1.0."""
        header_str = LINE_SEP.join(self.keyvalue_v110str_ls) + LINE_SEP + \
            LINE_TERMINATOR
        return header_str

    @property
    def keyvalue_v200_ls(self):
        """Return KeyValue list of strings following spec v2.0.0."""
        keyvalue = [self.timestamp] + [KEYVALUE_SEP_V200.join([k, v])
                                       for k, v in self.keyvalue_tuple_ls]
        return keyvalue

    @property
    def strv200(self):
        """Return header string following spec v2.0.0."""
        header_str = LINE_SEP.join(self.keyvalue_v200_ls) + LINE_SEP + \
            LINE_TERMINATOR
        return header_str

    @property
    def num_lines(self):
        return len(self.__str__().split(LINE_SEP))
juga's avatar
juga committed
199
200


juga's avatar
juga committed
201
class V3BWLine(object):
juga's avatar
juga committed
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
    """
    Create a Bandwidth List line following the spec version 1.1.0.

    :param str node_id:
    :param int bw:
    :param dict kwargs: extra headers. Currently supported:

        - nickname, str
        - master_key_ed25519, str
        - rtt, int
        - time, str
        - sucess, int
        - error_stream, int
        - error_circ, int
        - error_misc, int
    """
juga's avatar
juga committed
218
219
220
    def __init__(self, node_id, bw, **kwargs):
        assert isinstance(node_id, str)
        assert isinstance(bw, int)
221
        assert node_id.startswith('$')
juga's avatar
juga committed
222
223
224
225
226
        self.node_id = node_id
        self.bw = bw
        [setattr(self, k, v) for k, v in kwargs.items()
         if k in BW_EXTRA_ARG_KEYVALUES]

juga's avatar
juga committed
227
    def __str__(self):
juga's avatar
juga committed
228
229
        return self.bw_strv110

juga's avatar
juga committed
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
    @classmethod
    def from_results(cls, results):
        success_results = [r for r in results if isinstance(r, ResultSuccess)]
        # log.debug('len(success_results) %s', len(success_results))
        node_id = '$' + results[0].fingerprint
        bw = cls.bw_from_results(success_results)
        kwargs = dict()
        kwargs['nick'] = results[0].nickname
        if getattr(results[0], 'master_key_ed25519'):
            kwargs['master_key_ed25519'] = results[0].master_key_ed25519
        kwargs['rtt'] = cls.rtt_from_results(success_results)
        kwargs['time'] = cls.last_time_from_results(results)
        kwargs.update(cls.result_types_from_results(results))
        bwl = cls(node_id, bw, **kwargs)
        return bwl

    @classmethod
    def from_data(cls, data, fingerprint):
        assert fingerprint in data
        return cls.from_results(data[fingerprint])

juga's avatar
juga committed
251
252
253
254
255
256
    @classmethod
    def from_bw_line_v110(cls, line):
        assert isinstance(line, str)
        kwargs = dict([kv.split(KEYVALUE_SEP_V110)
                       for kv in line.split(BW_KEYVALUE_SEP_V110)
                       if kv.split(KEYVALUE_SEP_V110)[0] in BW_KEYVALUES])
juga's avatar
juga committed
257
258
259
        for k, v in kwargs.items():
            if k in BW_KEYVALUES_INT:
                kwargs[k] = int(v)
juga's avatar
juga committed
260
261
262
        bw_line = cls(**kwargs)
        return bw_line

juga's avatar
juga committed
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
    @staticmethod
    def last_time_from_results(results):
        return unixts_to_isodt_str(round(max([r.time for r in results])))

    @staticmethod
    def rtt_from_results(results):
        # convert from miliseconds to seconds
        rtts = [(round(rtt * 1000)) for r in results for rtt in r.rtts]
        rtt = round(median(rtts))
        return rtt

    @staticmethod
    def result_types_from_results(results):
        rt_dict = dict([(result_type_to_key(rt.value),
                         num_results_of_type(results, rt.value))
                        for rt in _ResultType])
        return rt_dict

juga's avatar
juga committed
281
282
283
284
    @staticmethod
    def bw_from_results(results):
        median_bw = median([dl['amount'] / dl['duration']
                            for r in results for dl in r.downloads])
285
286
287
288
289
290
291
292
293
294
295
        # If a relay has MaxAdvertisedBandwidth set, they may be capable of
        # some large amount of bandwidth but prefer if they didn't receive it.
        # We also could have managed to measure them faster than their
        # {,Relay}BandwidthRate somehow.
        #
        # See https://github.com/pastly/simple-bw-scanner/issues/155 and
        # https://trac.torproject.org/projects/tor/ticket/8494
        #
        # Note how this isn't some measured-by-us average of bandwidth. It's
        # the first value on the 'bandwidth' line in the relay's server
        # descriptor.
juga's avatar
juga committed
296
297
298
299
300
301
302
        bw = median_bw
        relay_average_bw = [r.relay_average_bandwidth for r in results
                            if r.relay_average_bandwidth is not None]
        if relay_average_bw:
            median_relay_average_bw = median(relay_average_bw)
            if median_bw > median_relay_average_bw:
                bw = median_relay_average_bw
juga's avatar
juga committed
303
        # convert to KB and ensure it's at least 1
304
        bw_kb = max(round(bw / 1024), 1)
juga's avatar
juga committed
305
306
        return bw_kb

juga's avatar
juga committed
307
308
309
310
311
312
313
    @property
    def bw_keyvalue_tuple_ls(self):
        """Return list of KeyValue Bandwidth Line tuples."""
        # sort the list to generate determinist headers
        keyvalue_tuple_ls = sorted([(k, v) for k, v in self.__dict__.items()
                                    if k in BW_KEYVALUES])
        return keyvalue_tuple_ls
juga's avatar
juga committed
314

juga's avatar
juga committed
315
316
317
318
319
320
321
322
    @property
    def bw_keyvalue_v110str_ls(self):
        """Return list of KeyValue Bandwidth Line strings following
        spec v1.1.0.
        """
        bw_keyvalue_str = [KEYVALUE_SEP_V110 .join([k, str(v)])
                           for k, v in self.bw_keyvalue_tuple_ls]
        return bw_keyvalue_str
juga's avatar
juga committed
323

juga's avatar
juga committed
324
325
326
327
328
329
330
331
332
333
334
    @property
    def bw_strv110(self):
        """Return Bandwidth Line string following spec v1.1.0."""
        bw_line_str = BW_KEYVALUE_SEP_V110.join(
                        self.bw_keyvalue_v110str_ls) + LINE_SEP
        if len(bw_line_str) > BW_LINE_SIZE:
            # if this is the case, probably there are too many KeyValues,
            # or the limit needs to be changed in Tor
            log.warn("The bandwidth line %s is longer than %s",
                     len(bw_line_str), BW_LINE_SIZE)
        return bw_line_str
juga's avatar
juga committed
335

juga's avatar
juga committed
336

337
class V3BWFile(object):
juga's avatar
juga committed
338
339
340
341
342
343
    """
    Create a Bandwidth List file following spec version 1.1.0

    :param V3BWHeader v3bwheader: header
    :param list v3bwlines: V3BWLines
    """
juga's avatar
juga committed
344
345
346
347
348
349
350
351
    def __init__(self, v3bwheader, v3bwlines):
        self.header = v3bwheader
        self.bw_lines = v3bwlines

    def __str__(self):
        return str(self.header) + ''.join([str(bw_line)
                                           for bw_line in self.bw_lines])

juga's avatar
juga committed
352
    @classmethod
juga's avatar
juga committed
353
354
    def from_results(cls, results, state_fpath='',
                     scale_constant=None):
juga's avatar
juga committed
355
356
        bw_lines = [V3BWLine.from_results(results[fp]) for fp in results]
        bw_lines = sorted(bw_lines, key=lambda d: d.bw, reverse=True)
juga's avatar
juga committed
357
358
359
360
        if scale_constant:
            bw_lines = cls.bw_sbws_scale(bw_lines, scale_constant)
            cls.warn_if_not_accurate_enough(bw_lines, scale_constant)
        header = V3BWHeader.from_results(results, state_fpath)
juga's avatar
juga committed
361
362
363
        f = cls(header, bw_lines)
        return f

juga's avatar
juga committed
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
    @staticmethod
    def bw_sbws_scale(bw_lines, scale_constant=SCALE_CONSTANT,
                      reverse=False):
        """Return a new V3BwLine list scaled using sbws method.

        :param list bw_lines:
            bw lines to scale, not self.bw_lines,
            since this method will be before self.bw_lines have been
            initialized.
        :param int scale_constant:
            the constant to multiply by the ratio and
            the bandwidth to obtain the new bandwidth
        :returns list: V3BwLine list
        """
        # If a relay has MaxAdvertisedBandwidth set, they may be capable of
        # some large amount of bandwidth but prefer if they didn't receive it.
        # We also could have managed to measure them faster than their
        # {,Relay}BandwidthRate somehow.
        #
        # See https://github.com/pastly/simple-bw-scanner/issues/155 and
        # https://trac.torproject.org/projects/tor/ticket/8494
        #
        # Note how this isn't some measured-by-us average of bandwidth. It's
        # the first value on the 'bandwidth' line in the relay's server
        # descriptor.
        log.debug('Scaling bandwidth using sbws method.')
        m = median([l.bw for l in bw_lines])
        bw_lines_scaled = copy.deepcopy(bw_lines)
        for l in bw_lines_scaled:
            # min is to limit the bw to descriptor average-bandwidth
            # max to avoid bandwidth with 0 value
            l.bw = max(round(min(l.desc_avg_bw_bs,
                                 l.bw_bs_median * scale_constant / m)
                             / 1000), 1)
        return sorted(bw_lines_scaled, key=lambda x: x.bw, reverse=reverse)

    @staticmethod
    def warn_if_not_accurate_enough(bw_lines,
                                    scale_constant=SCALE_CONSTANT):
        margin = 0.001
        accuracy_ratio = median([l.bw for l in bw_lines]) / scale_constant
        log.info('The generated lines are within {:.5}% of what they should '
                 'be'.format((1 - accuracy_ratio) * 100))
        if accuracy_ratio < 1 - margin or accuracy_ratio > 1 + margin:
            log.warning('There was %f%% error and only +/- %f%% is '
                        'allowed', (1 - accuracy_ratio) * 100, margin * 100)

juga's avatar
juga committed
411
    @property
juga's avatar
juga committed
412
413
    def sum_bw(self):
        return sum([l.bw for l in self.bw_lines])
juga's avatar
juga committed
414
415

    @property
juga's avatar
juga committed
416
    def num(self):
juga's avatar
juga committed
417
418
419
        return len(self.bw_lines)

    @property
juga's avatar
juga committed
420
421
422
423
424
425
    def mean_bw(self):
        return mean([l.bw for l in self.bw_lines])

    @property
    def median_bw(self):
        return median([l.bw for l in self.bw_lines])
juga's avatar
juga committed
426

juga's avatar
juga committed
427
428
429
430
431
    def write(self, output):
        if output == '/dev/stdout':
            log.info("Writing to stdout is not supported.")
            return
        log.info('Writing v3bw file to %s', output)
432
433
        # To avoid inconsistent reads, the bandwidth data is written to an
        # archive path, then atomically symlinked to 'latest.v3bw'
juga's avatar
juga committed
434
435
        out_dir = os.path.dirname(output)
        out_link = os.path.join(out_dir, 'latest.v3bw')
436
        out_link_tmp = out_link + '.tmp'
juga's avatar
juga committed
437
438
439
440
441
442
        with DirectoryLock(out_dir):
            with open(output, 'wt') as fd:
                fd.write(str(self.header))
                for line in self.bw_lines:
                    fd.write(str(line))
            output_basename = os.path.basename(output)
443
444
445
446
447
448
449
450
451
452
            # To atomically symlink a file, we need to create a temporary link,
            # then rename it to the final link name. (POSIX guarantees that
            # rename is atomic.)
            log.debug('Creating symlink {} -> {}.'
                      .format(out_link_tmp, output_basename))
            os.symlink(output_basename, out_link_tmp)
            log.debug('Renaming symlink {} -> {} to {} -> {}.'
                      .format(out_link_tmp, output_basename,
                              out_link, output_basename))
            os.rename(out_link_tmp, out_link)