v3bwfile.py 15.6 KB
Newer Older
juga's avatar
juga committed
1
2
3
4
5
# -*- coding: utf-8 -*-
"""Classes and functions that create the bandwidth measurements document
(v3bw) used by bandwidth authorities."""

import logging
juga's avatar
juga committed
6
import os
juga's avatar
juga committed
7
from statistics import median
juga's avatar
juga committed
8

9
from sbws import __version__
juga's avatar
juga committed
10
from sbws.globals import SPEC_VERSION, BW_LINE_SIZE
juga's avatar
juga committed
11
from sbws.lib.resultdump import ResultSuccess, _ResultType
12
from sbws.util.filelock import DirectoryLock
Matt Traudt's avatar
Matt Traudt committed
13
from sbws.util.timestamp import now_isodt_str, unixts_to_isodt_str
14
from sbws.util.state import State
juga's avatar
juga committed
15
16
17

log = logging.getLogger(__name__)

18
LINE_SEP = '\n'
19
20
21
22
23
24
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
25
UNORDERED_KEYVALUES = EXTRA_ARG_KEYVALUES + ['latest_bandwidth']
26
27
# List of all the KeyValues currently being used to generate the file
ALL_KEYVALUES = ['version'] + UNORDERED_KEYVALUES
28
TERMINATOR = '===='
juga's avatar
juga committed
29
30
# Num header lines in v1.1.0 using all the KeyValues
NUM_LINES_HEADER_V110 = len(ALL_KEYVALUES) + 2
31
32
LINE_TERMINATOR = TERMINATOR + LINE_SEP

juga's avatar
juga committed
33
34
# KeyValue separator in Bandwidth Lines
BW_KEYVALUE_SEP_V110 = ' '
juga's avatar
juga committed
35
BW_EXTRA_ARG_KEYVALUES = ['master_key_ed25519', 'nick', 'rtt', 'time',
juga's avatar
juga committed
36
                          'success', 'error_stream', 'error_circ',
juga's avatar
juga committed
37
                          'error_misc']
38
BW_KEYVALUES_INT = ['bw', 'rtt', 'success', 'error_stream',
juga's avatar
juga committed
39
                    'error_circ', 'error_misc']
juga's avatar
juga committed
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
BW_KEYVALUES = ['node_id', 'bw'] + BW_EXTRA_ARG_KEYVALUES


def total_bw(bw_lines):
    return sum([l.bw for l in bw_lines])


def avg_bw(bw_lines):
    assert len(bw_lines) > 0
    return total_bw(bw_lines) / len(bw_lines)


def scale_lines(bw_lines, scale_constant):
    avg = avg_bw(bw_lines)
    for line in bw_lines:
        line.bw = round(line.bw / avg * scale_constant)
    warn_if_not_accurate_enough(bw_lines, scale_constant)
    return bw_lines


def warn_if_not_accurate_enough(bw_lines, scale_constant):
    margin = 0.001
    accuracy_ratio = avg_bw(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
69

juga's avatar
juga committed
70
71
72
73
74
75
76
77
78
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('-', '_')


79
class V3BWHeader(object):
juga's avatar
juga committed
80
81
82
83
    """
    Create a bandwidth measurements (V3bw) header
    following bandwidth measurements document spec version 1.1.0.

84
    :param str timestamp: timestamp in Unix Epoch seconds of the most recent
85
        generator result.
juga's avatar
juga committed
86
87
88
    :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
89
    :param dict kwargs: extra headers. Currently supported:
juga's avatar
juga committed
90

91
92
93
94
        - 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
95
    """
96
    def __init__(self, timestamp, **kwargs):
juga's avatar
juga committed
97
98
99
100
        assert isinstance(timestamp, str)
        for v in kwargs.values():
            assert isinstance(v, str)
        self.timestamp = timestamp
101
102
103
104
        # 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__)
105
        self.file_created = kwargs.get('file_created', now_isodt_str())
juga's avatar
juga committed
106
        # latest_bandwidth should not be in kwargs, since it MUST be the
107
        # same as timestamp
juga's avatar
juga committed
108
        self.latest_bandwidth = unixts_to_isodt_str(timestamp)
109
110
        [setattr(self, k, v) for k, v in kwargs.items()
         if k in EXTRA_ARG_KEYVALUES]
juga's avatar
juga committed
111

112
113
114
115
116
    def __str__(self):
        if self.version == '1.1.0':
            return self.strv110
        return self.strv200

juga's avatar
juga committed
117
    @classmethod
juga's avatar
juga committed
118
    def from_results(cls, results, state_fpath=''):
juga's avatar
juga committed
119
120
121
        kwargs = dict()
        latest_bandwidth = cls.latest_bandwidth_from_results(results)
        earliest_bandwidth = cls.earliest_bandwidth_from_results(results)
juga's avatar
juga committed
122
        generator_started = cls.generator_started_from_file(state_fpath)
juga's avatar
juga committed
123
124
125
126
127
128
129
130
        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

131
132
133
134
    @classmethod
    def from_lines_v110(cls, lines):
        """
        :param list lines: list of lines to parse
135
        :returns: tuple of V3BWHeader object and non-header lines
136
137
138
139
        """
        assert isinstance(lines, list)
        try:
            index_terminator = lines.index(TERMINATOR)
Matt Traudt's avatar
Matt Traudt committed
140
        except ValueError:
141
142
143
            # is not a bw file or is v100
            log.warn('Terminator is not in lines')
            return None
144
        ts = lines[0]
145
        kwargs = dict([l.split(KEYVALUE_SEP_V110)
146
                       for l in lines[:index_terminator]
147
                       if l.split(KEYVALUE_SEP_V110)[0] in ALL_KEYVALUES])
148
        h = cls(ts, **kwargs)
149
150
       # last line is new line
        return h, lines[index_terminator + 1:-1]
151
152
153
154

    @classmethod
    def from_text_v110(self, text):
        """
155
        :param str text: text to parse
156
        :returns: tuple of V3BWHeader object and non-header lines
157
158
159
        """
        assert isinstance(text, str)
        return self.from_lines_v110(text.split(LINE_SEP))
juga's avatar
juga committed
160
161

    @staticmethod
juga's avatar
juga committed
162
    def generator_started_from_file(state_fpath):
163
164
165
166
        '''
        ISO formatted timestamp for the time when the scanner process most
        recently started.
        '''
juga's avatar
juga committed
167
        state = State(state_fpath)
168
169
170
171
        if 'scanner_started' in state:
            return state['scanner_started']
        else:
            return None
juga's avatar
juga committed
172
173

    @staticmethod
juga's avatar
juga committed
174
    def latest_bandwidth_from_results(results):
juga's avatar
juga committed
175
        return round(max([r.time for fp in results for r in results[fp]]))
juga's avatar
juga committed
176
177
178

    @staticmethod
    def earliest_bandwidth_from_results(results):
juga's avatar
juga committed
179
        return round(min([r.time for fp in results for r in results[fp]]))
juga's avatar
juga committed
180

juga's avatar
juga committed
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
    @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
225
226


juga's avatar
juga committed
227
class V3BWLine(object):
juga's avatar
juga committed
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
    """
    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
244
245
246
    def __init__(self, node_id, bw, **kwargs):
        assert isinstance(node_id, str)
        assert isinstance(bw, int)
247
        assert node_id.startswith('$')
juga's avatar
juga committed
248
249
250
251
252
        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
253
    def __str__(self):
juga's avatar
juga committed
254
255
        return self.bw_strv110

juga's avatar
juga committed
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
    @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
277
278
279
280
281
282
    @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
283
284
285
        for k, v in kwargs.items():
            if k in BW_KEYVALUES_INT:
                kwargs[k] = int(v)
juga's avatar
juga committed
286
287
288
        bw_line = cls(**kwargs)
        return bw_line

juga's avatar
juga committed
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
    @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
307
308
309
310
    @staticmethod
    def bw_from_results(results):
        median_bw = median([dl['amount'] / dl['duration']
                            for r in results for dl in r.downloads])
311
312
313
314
315
316
317
318
319
320
321
        # 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
322
323
324
325
326
327
328
        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
329
        # convert to KB and ensure it's at least 1
330
        bw_kb = max(round(bw / 1024), 1)
juga's avatar
juga committed
331
332
        return bw_kb

juga's avatar
juga committed
333
334
335
336
337
338
339
    @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
340

juga's avatar
juga committed
341
342
343
344
345
346
347
348
    @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
349

juga's avatar
juga committed
350
351
352
353
354
355
356
357
358
359
360
    @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
361

juga's avatar
juga committed
362

363
class V3BWFile(object):
juga's avatar
juga committed
364
365
366
367
368
369
    """
    Create a Bandwidth List file following spec version 1.1.0

    :param V3BWHeader v3bwheader: header
    :param list v3bwlines: V3BWLines
    """
juga's avatar
juga committed
370
371
372
373
374
375
376
377
    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
378
    @classmethod
juga's avatar
juga committed
379
380
    def from_results(cls, results, state_fpath='',
                     scale_constant=None):
juga's avatar
juga committed
381
382
        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
383
384
385
386
        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
387
388
389
        f = cls(header, bw_lines)
        return f

juga's avatar
juga committed
390
391
392
393
394
395
396
397
398
399
400
401
    @property
    def total_bw(self):
        return total_bw(self.bw_lines)

    @property
    def num_lines(self):
        return len(self.bw_lines)

    @property
    def avg_bw(self):
        return self.total_bw / self.num_lines

juga's avatar
juga committed
402
403
404
405
406
    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)
407
408
        # To avoid inconsistent reads, the bandwidth data is written to an
        # archive path, then atomically symlinked to 'latest.v3bw'
juga's avatar
juga committed
409
410
        out_dir = os.path.dirname(output)
        out_link = os.path.join(out_dir, 'latest.v3bw')
411
        out_link_tmp = out_link + '.tmp'
juga's avatar
juga committed
412
413
414
415
416
417
        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)
418
419
420
421
422
423
424
425
426
427
            # 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)