contrib/synthrepo.py
author pacien <pacien.trangirard@pacien.net>
Fri, 21 Apr 2023 14:32:58 +0200
changeset 50525 1a4f54574e3d
parent 48935 2cce2fa5bcf7
permissions -rw-r--r--
stabletailgraph: clarify naiveness of current implementation Both the naive and the actual versions of the algorithms are going to co-exist for the tests. This makes is clearer that this one is naive.
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
     1
# synthrepo.py - repo synthesis
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
     2
#
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
     3
# Copyright 2012 Facebook
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
     4
#
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
     5
# This software may be used and distributed according to the terms of the
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
     6
# GNU General Public License version 2 or any later version.
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
     7
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
     8
'''synthesize structurally interesting change history
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
     9
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    10
This extension is useful for creating a repository with properties
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    11
that are statistically similar to an existing repository. During
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    12
analysis, a simple probability table is constructed from the history
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    13
of an existing repository.  During synthesis, these properties are
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    14
reconstructed.
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    15
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    16
Properties that are analyzed and synthesized include the following:
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    17
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    18
- Lines added or removed when an existing file is modified
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    19
- Number and sizes of files added
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    20
- Number of files removed
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    21
- Line lengths
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    22
- Topological distance to parent changeset(s)
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    23
- Probability of a commit being a merge
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    24
- Probability of a newly added file being added to a new directory
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    25
- Interarrival time, and time zone, of commits
22709
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
    26
- Number of files in each directory
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    27
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    28
A few obvious properties that are not currently handled realistically:
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    29
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    30
- Merges are treated as regular commits with two parents, which is not
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    31
  realistic
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    32
- Modifications are not treated as operations on hunks of lines, but
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    33
  as insertions and deletions of randomly chosen single lines
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    34
- Committer ID (always random)
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    35
- Executability of files
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    36
- Symlinks and binary files are ignored
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    37
'''
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    38
28563
62250a48dc7f contrib: synthrepo use absolute_import
Pulkit Goyal <7895pulkit@gmail.com>
parents: 26587
diff changeset
    39
import bisect
62250a48dc7f contrib: synthrepo use absolute_import
Pulkit Goyal <7895pulkit@gmail.com>
parents: 26587
diff changeset
    40
import collections
62250a48dc7f contrib: synthrepo use absolute_import
Pulkit Goyal <7895pulkit@gmail.com>
parents: 26587
diff changeset
    41
import itertools
62250a48dc7f contrib: synthrepo use absolute_import
Pulkit Goyal <7895pulkit@gmail.com>
parents: 26587
diff changeset
    42
import json
62250a48dc7f contrib: synthrepo use absolute_import
Pulkit Goyal <7895pulkit@gmail.com>
parents: 26587
diff changeset
    43
import os
62250a48dc7f contrib: synthrepo use absolute_import
Pulkit Goyal <7895pulkit@gmail.com>
parents: 26587
diff changeset
    44
import random
62250a48dc7f contrib: synthrepo use absolute_import
Pulkit Goyal <7895pulkit@gmail.com>
parents: 26587
diff changeset
    45
import sys
62250a48dc7f contrib: synthrepo use absolute_import
Pulkit Goyal <7895pulkit@gmail.com>
parents: 26587
diff changeset
    46
import time
29205
a0939666b836 py3: move up symbol imports to enforce import-checker rules
Yuya Nishihara <yuya@tcha.org>
parents: 28563
diff changeset
    47
a0939666b836 py3: move up symbol imports to enforce import-checker rules
Yuya Nishihara <yuya@tcha.org>
parents: 28563
diff changeset
    48
from mercurial.i18n import _
a0939666b836 py3: move up symbol imports to enforce import-checker rules
Yuya Nishihara <yuya@tcha.org>
parents: 28563
diff changeset
    49
from mercurial.node import (
a0939666b836 py3: move up symbol imports to enforce import-checker rules
Yuya Nishihara <yuya@tcha.org>
parents: 28563
diff changeset
    50
    nullid,
a0939666b836 py3: move up symbol imports to enforce import-checker rules
Yuya Nishihara <yuya@tcha.org>
parents: 28563
diff changeset
    51
    nullrev,
a0939666b836 py3: move up symbol imports to enforce import-checker rules
Yuya Nishihara <yuya@tcha.org>
parents: 28563
diff changeset
    52
    short,
a0939666b836 py3: move up symbol imports to enforce import-checker rules
Yuya Nishihara <yuya@tcha.org>
parents: 28563
diff changeset
    53
)
28563
62250a48dc7f contrib: synthrepo use absolute_import
Pulkit Goyal <7895pulkit@gmail.com>
parents: 26587
diff changeset
    54
from mercurial import (
62250a48dc7f contrib: synthrepo use absolute_import
Pulkit Goyal <7895pulkit@gmail.com>
parents: 26587
diff changeset
    55
    context,
38588
1c93e0237a24 diffutil: move the module out of utils package
Yuya Nishihara <yuya@tcha.org>
parents: 38587
diff changeset
    56
    diffutil,
28563
62250a48dc7f contrib: synthrepo use absolute_import
Pulkit Goyal <7895pulkit@gmail.com>
parents: 26587
diff changeset
    57
    error,
62250a48dc7f contrib: synthrepo use absolute_import
Pulkit Goyal <7895pulkit@gmail.com>
parents: 26587
diff changeset
    58
    hg,
48116
5ced12cfa41b errors: raise InputError on bad revset to revrange() iff provided by the user
Martin von Zweigbergk <martinvonz@google.com>
parents: 45942
diff changeset
    59
    logcmdutil,
28563
62250a48dc7f contrib: synthrepo use absolute_import
Pulkit Goyal <7895pulkit@gmail.com>
parents: 26587
diff changeset
    60
    patch,
43004
c07812bdd568 synthrepo: use pycompat.xrange
Gregory Szorc <gregory.szorc@gmail.com>
parents: 41398
diff changeset
    61
    pycompat,
32337
46ba2cdda476 registrar: move cmdutil.command to registrar module (API)
Yuya Nishihara <yuya@tcha.org>
parents: 32291
diff changeset
    62
    registrar,
28563
62250a48dc7f contrib: synthrepo use absolute_import
Pulkit Goyal <7895pulkit@gmail.com>
parents: 26587
diff changeset
    63
)
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
    64
from mercurial.utils import dateutil
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    65
29841
d5883fd055c6 extensions: change magic "shipped with hg" string
Augie Fackler <augie@google.com>
parents: 29216
diff changeset
    66
# Note for extension authors: ONLY specify testedwith = 'ships-with-hg-core' for
25186
80c5b2666a96 extensions: document that `testedwith = 'internal'` is special
Augie Fackler <augie@google.com>
parents: 24306
diff changeset
    67
# extensions which SHIP WITH MERCURIAL. Non-mainline extensions should
80c5b2666a96 extensions: document that `testedwith = 'internal'` is special
Augie Fackler <augie@google.com>
parents: 24306
diff changeset
    68
# be specifying the version(s) of Mercurial they are tested with, or
80c5b2666a96 extensions: document that `testedwith = 'internal'` is special
Augie Fackler <augie@google.com>
parents: 24306
diff changeset
    69
# leave the attribute unspecified.
29841
d5883fd055c6 extensions: change magic "shipped with hg" string
Augie Fackler <augie@google.com>
parents: 29216
diff changeset
    70
testedwith = 'ships-with-hg-core'
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    71
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    72
cmdtable = {}
32337
46ba2cdda476 registrar: move cmdutil.command to registrar module (API)
Yuya Nishihara <yuya@tcha.org>
parents: 32291
diff changeset
    73
command = registrar.command(cmdtable)
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    74
32291
bd872f64a8ba cleanup: use set literals
Martin von Zweigbergk <martinvonz@google.com>
parents: 29841
diff changeset
    75
newfile = {'new fi', 'rename', 'copy f', 'copy t'}
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    76
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
    77
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    78
def zerodict():
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    79
    return collections.defaultdict(lambda: 0)
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    80
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
    81
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    82
def roundto(x, k):
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    83
    if x > k * 2:
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    84
        return int(round(x / float(k)) * k)
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    85
    return int(round(x))
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    86
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
    87
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    88
def parsegitdiff(lines):
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    89
    filename, mar, lineadd, lineremove = None, None, zerodict(), 0
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    90
    binary = False
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    91
    for line in lines:
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    92
        start = line[:6]
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    93
        if start == 'diff -':
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    94
            if filename:
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    95
                yield filename, mar, lineadd, lineremove, binary
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    96
            mar, lineadd, lineremove, binary = 'm', zerodict(), 0, False
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    97
            filename = patch.gitre.match(line).group(1)
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    98
        elif start in newfile:
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
    99
            mar = 'a'
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   100
        elif start == 'GIT bi':
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   101
            binary = True
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   102
        elif start == 'delete':
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   103
            mar = 'r'
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   104
        elif start:
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   105
            s = start[0]
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   106
            if s == '-' and not line.startswith('--- '):
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   107
                lineremove += 1
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   108
            elif s == '+' and not line.startswith('+++ '):
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   109
                lineadd[roundto(len(line) - 1, 5)] += 1
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   110
    if filename:
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   111
        yield filename, mar, lineadd, lineremove, binary
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   112
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   113
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   114
@command(
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   115
    'analyze',
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   116
    [
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   117
        ('o', 'output', '', _('write output to given file'), _('FILE')),
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   118
        ('r', 'rev', [], _('analyze specified revisions'), _('REV')),
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   119
    ],
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   120
    _('hg analyze'),
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   121
    optionalrepo=True,
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   122
)
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   123
def analyze(ui, repo, *revs, **opts):
45942
89a2afe31e82 formating: upgrade to black 20.8b1
Augie Fackler <raf@durin42.com>
parents: 43104
diff changeset
   124
    """create a simple model of a repository to use for later synthesis
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   125
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   126
    This command examines every changeset in the given range (or all
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   127
    of history if none are specified) and creates a simple statistical
22709
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   128
    model of the history of the repository. It also measures the directory
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   129
    structure of the repository as checked out.
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   130
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   131
    The model is written out to a JSON file, and can be used by
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   132
    :hg:`synthesize` to create or augment a repository with synthetic
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   133
    commits that have a structure that is statistically similar to the
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   134
    analyzed repository.
45942
89a2afe31e82 formating: upgrade to black 20.8b1
Augie Fackler <raf@durin42.com>
parents: 43104
diff changeset
   135
    """
22709
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   136
    root = repo.root
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   137
    if not root.endswith(os.path.sep):
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   138
        root += os.path.sep
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   139
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   140
    revs = list(revs)
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   141
    revs.extend(opts['rev'])
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   142
    if not revs:
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   143
        revs = [':']
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   144
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   145
    output = opts['output']
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   146
    if not output:
22709
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   147
        output = os.path.basename(root) + '.json'
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   148
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   149
    if output == '-':
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   150
        fp = sys.stdout
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   151
    else:
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   152
        fp = open(output, 'w')
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   153
22709
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   154
    # Always obtain file counts of each directory in the given root directory.
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   155
    def onerror(e):
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   156
        ui.warn(_('error walking directory structure: %s\n') % e)
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   157
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   158
    dirs = {}
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   159
    rootprefixlen = len(root)
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   160
    for dirpath, dirnames, filenames in os.walk(root, onerror=onerror):
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   161
        dirpathfromroot = dirpath[rootprefixlen:]
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   162
        dirs[dirpathfromroot] = len(filenames)
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   163
        if '.hg' in dirnames:
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   164
            dirnames.remove('.hg')
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   165
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   166
    lineschanged = zerodict()
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   167
    children = zerodict()
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   168
    p1distance = zerodict()
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   169
    p2distance = zerodict()
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   170
    linesinfilesadded = zerodict()
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   171
    fileschanged = zerodict()
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   172
    filesadded = zerodict()
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   173
    filesremoved = zerodict()
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   174
    linelengths = zerodict()
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   175
    interarrival = zerodict()
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   176
    parents = zerodict()
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   177
    dirsadded = zerodict()
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   178
    tzoffset = zerodict()
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   179
22709
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   180
    # If a mercurial repo is available, also model the commit history.
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   181
    if repo:
48116
5ced12cfa41b errors: raise InputError on bad revset to revrange() iff provided by the user
Martin von Zweigbergk <martinvonz@google.com>
parents: 45942
diff changeset
   182
        revs = logcmdutil.revrange(repo, revs)
22709
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   183
        revs.sort()
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   184
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   185
        progress = ui.makeprogress(
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   186
            _('analyzing'), unit=_('changesets'), total=len(revs)
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   187
        )
22709
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   188
        for i, rev in enumerate(revs):
38408
6540333acb95 synthrepo: use progress helper
Martin von Zweigbergk <martinvonz@google.com>
parents: 36607
diff changeset
   189
            progress.update(i)
22709
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   190
            ctx = repo[rev]
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   191
            pl = ctx.parents()
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   192
            pctx = pl[0]
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   193
            prev = pctx.rev()
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   194
            children[prev] += 1
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   195
            p1distance[rev - prev] += 1
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   196
            parents[len(pl)] += 1
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   197
            tzoffset[ctx.date()[1]] += 1
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   198
            if len(pl) > 1:
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   199
                p2distance[rev - pl[1].rev()] += 1
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   200
            if prev == rev - 1:
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   201
                lastctx = pctx
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   202
            else:
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   203
                lastctx = repo[rev - 1]
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   204
            if lastctx.rev() != nullrev:
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   205
                timedelta = ctx.date()[0] - lastctx.date()[0]
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   206
                interarrival[roundto(timedelta, 300)] += 1
38587
b62000a28812 diffutil: remove diffopts() in favor of diffallopts()
Yuya Nishihara <yuya@tcha.org>
parents: 38584
diff changeset
   207
            diffopts = diffutil.diffallopts(ui, {'git': True})
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   208
            diff = sum(
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   209
                (d.splitlines() for d in ctx.diff(pctx, opts=diffopts)), []
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   210
            )
22709
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   211
            fileadds, diradds, fileremoves, filechanges = 0, 0, 0, 0
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   212
            for filename, mar, lineadd, lineremove, isbin in parsegitdiff(diff):
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   213
                if isbin:
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   214
                    continue
48935
2cce2fa5bcf7 py3: replace pycompat.itervalues(x) with x.values()
Gregory Szorc <gregory.szorc@gmail.com>
parents: 48875
diff changeset
   215
                added = sum(lineadd.values(), 0)
22709
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   216
                if mar == 'm':
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   217
                    if added and lineremove:
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   218
                        lineschanged[
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   219
                            roundto(added, 5), roundto(lineremove, 5)
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   220
                        ] += 1
22709
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   221
                        filechanges += 1
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   222
                elif mar == 'a':
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   223
                    fileadds += 1
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   224
                    if '/' in filename:
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   225
                        filedir = filename.rsplit('/', 1)[0]
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   226
                        if filedir not in pctx.dirs():
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   227
                            diradds += 1
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   228
                    linesinfilesadded[roundto(added, 5)] += 1
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   229
                elif mar == 'r':
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   230
                    fileremoves += 1
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   231
                for length, count in lineadd.iteritems():
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   232
                    linelengths[length] += count
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   233
            fileschanged[filechanges] += 1
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   234
            filesadded[fileadds] += 1
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   235
            dirsadded[diradds] += 1
889789a2ca9f contrib/synthrepo: walk a repo's directory structure during analysis
Mike Edgar <adgar@google.com>
parents: 22708
diff changeset
   236
            filesremoved[fileremoves] += 1
38409
ce65c25dc161 synthrepo: close progress topics
Martin von Zweigbergk <martinvonz@google.com>
parents: 38408
diff changeset
   237
        progress.complete()
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   238
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   239
    invchildren = zerodict()
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   240
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   241
    for rev, count in children.iteritems():
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   242
        invchildren[count] += 1
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   243
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   244
    if output != '-':
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   245
        ui.status(_('writing output to %s\n') % output)
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   246
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   247
    def pronk(d):
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   248
        return sorted(d.iteritems(), key=lambda x: x[1], reverse=True)
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   249
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   250
    json.dump(
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   251
        {
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   252
            'revs': len(revs),
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   253
            'initdirs': pronk(dirs),
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   254
            'lineschanged': pronk(lineschanged),
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   255
            'children': pronk(invchildren),
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   256
            'fileschanged': pronk(fileschanged),
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   257
            'filesadded': pronk(filesadded),
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   258
            'linesinfilesadded': pronk(linesinfilesadded),
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   259
            'dirsadded': pronk(dirsadded),
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   260
            'filesremoved': pronk(filesremoved),
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   261
            'linelengths': pronk(linelengths),
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   262
            'parents': pronk(parents),
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   263
            'p1distance': pronk(p1distance),
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   264
            'p2distance': pronk(p2distance),
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   265
            'interarrival': pronk(interarrival),
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   266
            'tzoffset': pronk(tzoffset),
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   267
        },
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   268
        fp,
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   269
    )
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   270
    fp.close()
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   271
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   272
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   273
@command(
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   274
    'synthesize',
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   275
    [
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   276
        ('c', 'count', 0, _('create given number of commits'), _('COUNT')),
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   277
        ('', 'dict', '', _('path to a dictionary of words'), _('FILE')),
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   278
        ('', 'initfiles', 0, _('initial file count to create'), _('COUNT')),
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   279
    ],
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   280
    _('hg synthesize [OPTION].. DESCFILE'),
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   281
)
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   282
def synthesize(ui, repo, descpath, **opts):
45942
89a2afe31e82 formating: upgrade to black 20.8b1
Augie Fackler <raf@durin42.com>
parents: 43104
diff changeset
   283
    """synthesize commits based on a model of an existing repository
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   284
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   285
    The model must have been generated by :hg:`analyze`. Commits will
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   286
    be generated randomly according to the probabilities described in
22708
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   287
    the model. If --initfiles is set, the repository will be seeded with
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   288
    the given number files following the modeled repository's directory
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   289
    structure.
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   290
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   291
    When synthesizing new content, commit descriptions, and user
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   292
    names, words will be chosen randomly from a dictionary that is
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   293
    presumed to contain one word per line. Use --dict to specify the
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   294
    path to an alternate dictionary to use.
45942
89a2afe31e82 formating: upgrade to black 20.8b1
Augie Fackler <raf@durin42.com>
parents: 43104
diff changeset
   295
    """
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   296
    try:
17887
0e2846b2482c url: use open and not url.open for local files (issue3624)
Siddharth Agarwal <sid0@fb.com>
parents: 17734
diff changeset
   297
        fp = hg.openpath(ui, descpath)
25660
328739ea70c3 global: mass rewrite to use modern exception syntax
Gregory Szorc <gregory.szorc@gmail.com>
parents: 25186
diff changeset
   298
    except Exception as err:
26587
56b2bcea2529 error: get Abort from 'error' instead of 'util'
Pierre-Yves David <pierre-yves.david@fb.com>
parents: 25660
diff changeset
   299
        raise error.Abort('%s: %s' % (descpath, err[0].strerror))
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   300
    desc = json.load(fp)
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   301
    fp.close()
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   302
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   303
    def cdf(l):
18047
9196638b08ce synthrepo: do not crash if a list is empty
Bryan O'Sullivan <bryano@fb.com>
parents: 17887
diff changeset
   304
        if not l:
9196638b08ce synthrepo: do not crash if a list is empty
Bryan O'Sullivan <bryano@fb.com>
parents: 17887
diff changeset
   305
            return [], []
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   306
        vals, probs = zip(*sorted(l, key=lambda x: x[1], reverse=True))
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   307
        t = float(sum(probs, 0))
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   308
        s, cdfs = 0, []
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   309
        for v in probs:
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   310
            s += v
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   311
            cdfs.append(s / t)
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   312
        return vals, cdfs
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   313
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   314
    lineschanged = cdf(desc['lineschanged'])
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   315
    fileschanged = cdf(desc['fileschanged'])
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   316
    filesadded = cdf(desc['filesadded'])
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   317
    dirsadded = cdf(desc['dirsadded'])
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   318
    filesremoved = cdf(desc['filesremoved'])
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   319
    linelengths = cdf(desc['linelengths'])
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   320
    parents = cdf(desc['parents'])
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   321
    p1distance = cdf(desc['p1distance'])
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   322
    p2distance = cdf(desc['p2distance'])
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   323
    interarrival = cdf(desc['interarrival'])
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   324
    linesinfilesadded = cdf(desc['linesinfilesadded'])
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   325
    tzoffset = cdf(desc['tzoffset'])
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   326
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   327
    dictfile = opts.get('dict') or '/usr/share/dict/words'
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   328
    try:
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   329
        fp = open(dictfile, 'rU')
25660
328739ea70c3 global: mass rewrite to use modern exception syntax
Gregory Szorc <gregory.szorc@gmail.com>
parents: 25186
diff changeset
   330
    except IOError as err:
26587
56b2bcea2529 error: get Abort from 'error' instead of 'util'
Pierre-Yves David <pierre-yves.david@fb.com>
parents: 25660
diff changeset
   331
        raise error.Abort('%s: %s' % (dictfile, err.strerror))
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   332
    words = fp.read().splitlines()
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   333
    fp.close()
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   334
22708
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   335
    initdirs = {}
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   336
    if desc['initdirs']:
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   337
        for k, v in desc['initdirs']:
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   338
            initdirs[k.encode('utf-8').replace('.hg', '_hg')] = v
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   339
        initdirs = renamedirs(initdirs, words)
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   340
    initdirscdf = cdf(initdirs)
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   341
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   342
    def pick(cdf):
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   343
        return cdf[0][bisect.bisect_left(cdf[1], random.random())]
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   344
22708
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   345
    def pickpath():
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   346
        return os.path.join(pick(initdirscdf), random.choice(words))
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   347
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   348
    def makeline(minimum=0):
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   349
        total = max(minimum, pick(linelengths))
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   350
        c, l = 0, []
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   351
        while c < total:
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   352
            w = random.choice(words)
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   353
            c += len(w) + 1
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   354
            l.append(w)
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   355
        return ' '.join(l)
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   356
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   357
    wlock = repo.wlock()
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   358
    lock = repo.lock()
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   359
32291
bd872f64a8ba cleanup: use set literals
Martin von Zweigbergk <martinvonz@google.com>
parents: 29841
diff changeset
   360
    nevertouch = {'.hgsub', '.hgignore', '.hgtags'}
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   361
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   362
    _synthesizing = _('synthesizing')
22708
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   363
    _files = _('initial files')
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   364
    _changesets = _('changesets')
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   365
22708
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   366
    # Synthesize a single initial revision adding files to the repo according
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   367
    # to the modeled directory structure.
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   368
    initcount = int(opts['initfiles'])
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   369
    if initcount and initdirs:
41398
2ff8994ac71d cleanup: use repo['.'] instead of repo[None].p1()
Martin von Zweigbergk <martinvonz@google.com>
parents: 41397
diff changeset
   370
        pctx = repo['.']
23778
a5dbec255f14 synthrepo: new filenames must not also be new directories, and vice-versa
Mike Edgar <adgar@google.com>
parents: 23235
diff changeset
   371
        dirs = set(pctx.dirs())
22708
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   372
        files = {}
23778
a5dbec255f14 synthrepo: new filenames must not also be new directories, and vice-versa
Mike Edgar <adgar@google.com>
parents: 23235
diff changeset
   373
a5dbec255f14 synthrepo: new filenames must not also be new directories, and vice-versa
Mike Edgar <adgar@google.com>
parents: 23235
diff changeset
   374
        def validpath(path):
a5dbec255f14 synthrepo: new filenames must not also be new directories, and vice-versa
Mike Edgar <adgar@google.com>
parents: 23235
diff changeset
   375
            # Don't pick filenames which are already directory names.
a5dbec255f14 synthrepo: new filenames must not also be new directories, and vice-versa
Mike Edgar <adgar@google.com>
parents: 23235
diff changeset
   376
            if path in dirs:
a5dbec255f14 synthrepo: new filenames must not also be new directories, and vice-versa
Mike Edgar <adgar@google.com>
parents: 23235
diff changeset
   377
                return False
a5dbec255f14 synthrepo: new filenames must not also be new directories, and vice-versa
Mike Edgar <adgar@google.com>
parents: 23235
diff changeset
   378
            # Don't pick directories which were used as file names.
a5dbec255f14 synthrepo: new filenames must not also be new directories, and vice-versa
Mike Edgar <adgar@google.com>
parents: 23235
diff changeset
   379
            while path:
a5dbec255f14 synthrepo: new filenames must not also be new directories, and vice-versa
Mike Edgar <adgar@google.com>
parents: 23235
diff changeset
   380
                if path in files:
a5dbec255f14 synthrepo: new filenames must not also be new directories, and vice-versa
Mike Edgar <adgar@google.com>
parents: 23235
diff changeset
   381
                    return False
a5dbec255f14 synthrepo: new filenames must not also be new directories, and vice-versa
Mike Edgar <adgar@google.com>
parents: 23235
diff changeset
   382
                path = os.path.dirname(path)
a5dbec255f14 synthrepo: new filenames must not also be new directories, and vice-versa
Mike Edgar <adgar@google.com>
parents: 23235
diff changeset
   383
            return True
a5dbec255f14 synthrepo: new filenames must not also be new directories, and vice-versa
Mike Edgar <adgar@google.com>
parents: 23235
diff changeset
   384
38408
6540333acb95 synthrepo: use progress helper
Martin von Zweigbergk <martinvonz@google.com>
parents: 36607
diff changeset
   385
        progress = ui.makeprogress(_synthesizing, unit=_files, total=initcount)
43004
c07812bdd568 synthrepo: use pycompat.xrange
Gregory Szorc <gregory.szorc@gmail.com>
parents: 41398
diff changeset
   386
        for i in pycompat.xrange(0, initcount):
38408
6540333acb95 synthrepo: use progress helper
Martin von Zweigbergk <martinvonz@google.com>
parents: 36607
diff changeset
   387
            progress.update(i)
22708
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   388
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   389
            path = pickpath()
23778
a5dbec255f14 synthrepo: new filenames must not also be new directories, and vice-versa
Mike Edgar <adgar@google.com>
parents: 23235
diff changeset
   390
            while not validpath(path):
22708
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   391
                path = pickpath()
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   392
            data = '%s contents\n' % path
35398
2123e7629ec0 synthrepo: create filectx instance in 'filectxfn' callback
Martin von Zweigbergk <martinvonz@google.com>
parents: 34023
diff changeset
   393
            files[path] = data
23778
a5dbec255f14 synthrepo: new filenames must not also be new directories, and vice-versa
Mike Edgar <adgar@google.com>
parents: 23235
diff changeset
   394
            dir = os.path.dirname(path)
a5dbec255f14 synthrepo: new filenames must not also be new directories, and vice-versa
Mike Edgar <adgar@google.com>
parents: 23235
diff changeset
   395
            while dir and dir not in dirs:
a5dbec255f14 synthrepo: new filenames must not also be new directories, and vice-versa
Mike Edgar <adgar@google.com>
parents: 23235
diff changeset
   396
                dirs.add(dir)
a5dbec255f14 synthrepo: new filenames must not also be new directories, and vice-versa
Mike Edgar <adgar@google.com>
parents: 23235
diff changeset
   397
                dir = os.path.dirname(dir)
22708
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   398
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   399
        def filectxfn(repo, memctx, path):
35400
8a0cac20a1ad memfilectx: make changectx argument mandatory in constructor (API)
Martin von Zweigbergk <martinvonz@google.com>
parents: 35398
diff changeset
   400
            return context.memfilectx(repo, memctx, path, files[path])
22708
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   401
38408
6540333acb95 synthrepo: use progress helper
Martin von Zweigbergk <martinvonz@google.com>
parents: 36607
diff changeset
   402
        progress.complete()
22708
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   403
        message = 'synthesized wide repo with %d files' % (len(files),)
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   404
        mc = context.memctx(
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   405
            repo,
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   406
            [pctx.node(), nullid],
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   407
            message,
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   408
            files,
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   409
            filectxfn,
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   410
            ui.username(),
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   411
            '%d %d' % dateutil.makedate(),
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   412
        )
22708
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   413
        initnode = mc.commit()
24306
6ddc86eedc3b style: kill ersatz if-else ternary operators
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 23778
diff changeset
   414
        if ui.debugflag:
6ddc86eedc3b style: kill ersatz if-else ternary operators
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 23778
diff changeset
   415
            hexfn = hex
6ddc86eedc3b style: kill ersatz if-else ternary operators
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 23778
diff changeset
   416
        else:
6ddc86eedc3b style: kill ersatz if-else ternary operators
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 23778
diff changeset
   417
            hexfn = short
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   418
        ui.status(
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   419
            _('added commit %s with %d files\n') % (hexfn(initnode), len(files))
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   420
        )
22708
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   421
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   422
    # Synthesize incremental revisions to the repository, adding repo depth.
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   423
    count = int(opts['count'])
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   424
    heads = set(map(repo.changelog.rev, repo.heads()))
38408
6540333acb95 synthrepo: use progress helper
Martin von Zweigbergk <martinvonz@google.com>
parents: 36607
diff changeset
   425
    progress = ui.makeprogress(_synthesizing, unit=_changesets, total=count)
43004
c07812bdd568 synthrepo: use pycompat.xrange
Gregory Szorc <gregory.szorc@gmail.com>
parents: 41398
diff changeset
   426
    for i in pycompat.xrange(count):
38408
6540333acb95 synthrepo: use progress helper
Martin von Zweigbergk <martinvonz@google.com>
parents: 36607
diff changeset
   427
        progress.update(i)
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   428
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   429
        node = repo.changelog.node
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   430
        revs = len(repo)
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   431
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   432
        def pickhead(heads, distance):
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   433
            if heads:
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   434
                lheads = sorted(heads)
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   435
                rev = revs - min(pick(distance), revs)
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   436
                if rev < lheads[-1]:
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   437
                    rev = lheads[bisect.bisect_left(lheads, rev)]
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   438
                else:
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   439
                    rev = lheads[-1]
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   440
                return rev, node(rev)
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   441
            return nullrev, nullid
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   442
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   443
        r1 = revs - min(pick(p1distance), revs)
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   444
        p1 = node(r1)
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   445
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   446
        # the number of heads will grow without bound if we use a pure
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   447
        # model, so artificially constrain their proliferation
22472
2e2577b0ccbe contrib/synthrepo: only generate 2 parents if model contains merges
Mike Edgar <adgar@google.com>
parents: 22446
diff changeset
   448
        toomanyheads = len(heads) > random.randint(1, 20)
2e2577b0ccbe contrib/synthrepo: only generate 2 parents if model contains merges
Mike Edgar <adgar@google.com>
parents: 22446
diff changeset
   449
        if p2distance[0] and (pick(parents) == 2 or toomanyheads):
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   450
            r2, p2 = pickhead(heads.difference([r1]), p2distance)
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   451
        else:
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   452
            r2, p2 = nullrev, nullid
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   453
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   454
        pl = [p1, p2]
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   455
        pctx = repo[r1]
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   456
        mf = pctx.manifest()
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   457
        mfk = mf.keys()
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   458
        changes = {}
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   459
        if mfk:
43004
c07812bdd568 synthrepo: use pycompat.xrange
Gregory Szorc <gregory.szorc@gmail.com>
parents: 41398
diff changeset
   460
            for __ in pycompat.xrange(pick(fileschanged)):
c07812bdd568 synthrepo: use pycompat.xrange
Gregory Szorc <gregory.szorc@gmail.com>
parents: 41398
diff changeset
   461
                for __ in pycompat.xrange(10):
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   462
                    fctx = pctx.filectx(random.choice(mfk))
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   463
                    path = fctx.path()
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   464
                    if not (
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   465
                        path in nevertouch
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   466
                        or fctx.isbinary()
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   467
                        or 'l' in fctx.flags()
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   468
                    ):
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   469
                        break
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   470
                lines = fctx.data().splitlines()
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   471
                add, remove = pick(lineschanged)
43004
c07812bdd568 synthrepo: use pycompat.xrange
Gregory Szorc <gregory.szorc@gmail.com>
parents: 41398
diff changeset
   472
                for __ in pycompat.xrange(remove):
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   473
                    if not lines:
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   474
                        break
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   475
                    del lines[random.randrange(0, len(lines))]
43004
c07812bdd568 synthrepo: use pycompat.xrange
Gregory Szorc <gregory.szorc@gmail.com>
parents: 41398
diff changeset
   476
                for __ in pycompat.xrange(add):
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   477
                    lines.insert(random.randint(0, len(lines)), makeline())
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   478
                path = fctx.path()
35398
2123e7629ec0 synthrepo: create filectx instance in 'filectxfn' callback
Martin von Zweigbergk <martinvonz@google.com>
parents: 34023
diff changeset
   479
                changes[path] = '\n'.join(lines) + '\n'
43004
c07812bdd568 synthrepo: use pycompat.xrange
Gregory Szorc <gregory.szorc@gmail.com>
parents: 41398
diff changeset
   480
            for __ in pycompat.xrange(pick(filesremoved)):
c07812bdd568 synthrepo: use pycompat.xrange
Gregory Szorc <gregory.szorc@gmail.com>
parents: 41398
diff changeset
   481
                for __ in pycompat.xrange(10):
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   482
                    path = random.choice(mfk)
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   483
                    if path not in changes:
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   484
                        break
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   485
        if filesadded:
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   486
            dirs = list(pctx.dirs())
23235
4cdc3e2810b9 synthrepo: when adding files, ensure new path is not a directory
Mike Edgar <adgar@google.com>
parents: 23234
diff changeset
   487
            dirs.insert(0, '')
43004
c07812bdd568 synthrepo: use pycompat.xrange
Gregory Szorc <gregory.szorc@gmail.com>
parents: 41398
diff changeset
   488
        for __ in pycompat.xrange(pick(filesadded)):
23235
4cdc3e2810b9 synthrepo: when adding files, ensure new path is not a directory
Mike Edgar <adgar@google.com>
parents: 23234
diff changeset
   489
            pathstr = ''
4cdc3e2810b9 synthrepo: when adding files, ensure new path is not a directory
Mike Edgar <adgar@google.com>
parents: 23234
diff changeset
   490
            while pathstr in dirs:
4cdc3e2810b9 synthrepo: when adding files, ensure new path is not a directory
Mike Edgar <adgar@google.com>
parents: 23234
diff changeset
   491
                path = [random.choice(dirs)]
4cdc3e2810b9 synthrepo: when adding files, ensure new path is not a directory
Mike Edgar <adgar@google.com>
parents: 23234
diff changeset
   492
                if pick(dirsadded):
4cdc3e2810b9 synthrepo: when adding files, ensure new path is not a directory
Mike Edgar <adgar@google.com>
parents: 23234
diff changeset
   493
                    path.append(random.choice(words))
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   494
                path.append(random.choice(words))
23235
4cdc3e2810b9 synthrepo: when adding files, ensure new path is not a directory
Mike Edgar <adgar@google.com>
parents: 23234
diff changeset
   495
                pathstr = '/'.join(filter(None, path))
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   496
            data = (
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   497
                '\n'.join(
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   498
                    makeline()
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   499
                    for __ in pycompat.xrange(pick(linesinfilesadded))
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   500
                )
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   501
                + '\n'
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   502
            )
35398
2123e7629ec0 synthrepo: create filectx instance in 'filectxfn' callback
Martin von Zweigbergk <martinvonz@google.com>
parents: 34023
diff changeset
   503
            changes[pathstr] = data
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   504
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   505
        def filectxfn(repo, memctx, path):
35398
2123e7629ec0 synthrepo: create filectx instance in 'filectxfn' callback
Martin von Zweigbergk <martinvonz@google.com>
parents: 34023
diff changeset
   506
            if path not in changes:
2123e7629ec0 synthrepo: create filectx instance in 'filectxfn' callback
Martin von Zweigbergk <martinvonz@google.com>
parents: 34023
diff changeset
   507
                return None
35400
8a0cac20a1ad memfilectx: make changectx argument mandatory in constructor (API)
Martin von Zweigbergk <martinvonz@google.com>
parents: 35398
diff changeset
   508
            return context.memfilectx(repo, memctx, path, changes[path])
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   509
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   510
        if not changes:
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   511
            continue
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   512
        if revs:
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   513
            date = repo['tip'].date()[0] + pick(interarrival)
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   514
        else:
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   515
            date = time.time() - (86400 * count)
23234
944d6cfbe166 synthrepo: synthesized dates must be positive, fit in 32-bit signed ints
Mike Edgar <adgar@google.com>
parents: 22709
diff changeset
   516
        # dates in mercurial must be positive, fit in 32-bit signed integers.
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   517
        date = min(0x7FFFFFFF, max(0, date))
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   518
        user = random.choice(words) + '@' + random.choice(words)
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   519
        mc = context.memctx(
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   520
            repo,
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   521
            pl,
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   522
            makeline(minimum=2),
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   523
            sorted(changes),
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   524
            filectxfn,
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   525
            user,
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   526
            '%d %d' % (date, pick(tzoffset)),
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   527
        )
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   528
        newnode = mc.commit()
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   529
        heads.add(repo.changelog.rev(newnode))
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   530
        heads.discard(r1)
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   531
        heads.discard(r2)
38409
ce65c25dc161 synthrepo: close progress topics
Martin von Zweigbergk <martinvonz@google.com>
parents: 38408
diff changeset
   532
    progress.complete()
17734
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   533
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   534
    lock.release()
619068c280fd contrib: add a commit synthesizer for reproducing scaling problems
Bryan O'Sullivan <bryano@fb.com>
parents:
diff changeset
   535
    wlock.release()
22708
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   536
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   537
22708
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   538
def renamedirs(dirs, words):
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   539
    '''Randomly rename the directory names in the per-dir file count dict.'''
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   540
    wordgen = itertools.cycle(words)
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   541
    replacements = {'': ''}
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   542
22708
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   543
    def rename(dirpath):
45942
89a2afe31e82 formating: upgrade to black 20.8b1
Augie Fackler <raf@durin42.com>
parents: 43104
diff changeset
   544
        """Recursively rename the directory and all path prefixes.
22708
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   545
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   546
        The mapping from path to renamed path is stored for all path prefixes
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   547
        as in dynamic programming, ensuring linear runtime and consistent
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   548
        renaming regardless of iteration order through the model.
45942
89a2afe31e82 formating: upgrade to black 20.8b1
Augie Fackler <raf@durin42.com>
parents: 43104
diff changeset
   549
        """
22708
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   550
        if dirpath in replacements:
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   551
            return replacements[dirpath]
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   552
        head, _ = os.path.split(dirpath)
24306
6ddc86eedc3b style: kill ersatz if-else ternary operators
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 23778
diff changeset
   553
        if head:
6ddc86eedc3b style: kill ersatz if-else ternary operators
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 23778
diff changeset
   554
            head = rename(head)
6ddc86eedc3b style: kill ersatz if-else ternary operators
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 23778
diff changeset
   555
        else:
6ddc86eedc3b style: kill ersatz if-else ternary operators
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 23778
diff changeset
   556
            head = ''
29216
ead25aa27a43 py3: convert to next() function
timeless <timeless@mozdev.org>
parents: 29205
diff changeset
   557
        renamed = os.path.join(head, next(wordgen))
22708
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   558
        replacements[dirpath] = renamed
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   559
        return renamed
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 43004
diff changeset
   560
22708
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   561
    result = []
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   562
    for dirpath, count in dirs.iteritems():
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   563
        result.append([rename(dirpath.lstrip(os.sep)), count])
4c66e70c3488 contrib/synthrepo: generate initial repo contents using directory shape model
Mike Edgar <adgar@google.com>
parents: 22473
diff changeset
   564
    return result