python3-pytest-benchmark 3.0.0-1 / usr / lib / python3 / dist-packages / pytest_benchmark / plugin.py

from __future__ import division
from __future__ import print_function

import argparse
import gc
import json
import operator
import os
import platform
import sys
import time
import traceback
from collections import defaultdict
from datetime import datetime
from math import ceil
from math import isinf

import py
import pytest

from . import __version__
from .compat import INT
from .compat import XRANGE
from .timers import compute_timer_precision
from .timers import default_timer
from .utils import NameWrapper
from .utils import SecondsDecimal
from .utils import cached_property
from .utils import first_or_value
from .utils import format_dict
from .utils import format_time
from .utils import get_commit_info
from .utils import get_current_time
from .utils import get_tag
from .utils import load_timer
from .utils import parse_compare_fail
from .utils import parse_rounds
from .utils import parse_save
from .utils import parse_seconds
from .utils import parse_sort
from .utils import parse_timer
from .utils import parse_warmup
from .utils import report_progress
from .utils import time_unit

try:
    import statistics
except (ImportError, SyntaxError):
    statistics = False
    statistics_error = traceback.format_exc()
else:
    from .stats import Stats

NUMBER_FMT = "{0:,.4f}" if sys.version_info[:2] > (2, 6) else "{0:.4f}"
ALIGNED_NUMBER_FMT = "{0:>{1},.4f}{2:<{3}}" if sys.version_info[:2] > (2, 6) else "{0:>{1}.4f}{2:<{3}}"
HISTOGRAM_CURRENT = "now"


class PerformanceRegression(pytest.UsageError):
    pass


class FixtureAlreadyUsed(Exception):
    pass


def pytest_report_header(config):
    bs = config._benchmarksession

    return ("benchmark: %(version)s (defaults:"
            " timer=%(timer)s"
            " disable_gc=%(disable_gc)s"
            " min_rounds=%(min_rounds)s"
            " min_time=%(min_time)s"
            " max_time=%(max_time)s"
            " calibration_precision=%(calibration_precision)s"
            " warmup=%(warmup)s"
            " warmup_iterations=%(warmup_iterations)s"
            ")") % dict(
        bs.options,
        version=__version__,
        timer=bs.options.get("timer"),
    )


def pytest_addoption(parser):
    group = parser.getgroup("benchmark")
    group.addoption(
        "--benchmark-min-time",
        metavar="SECONDS", type=parse_seconds, default="0.000005",
        help="Minimum time per round in seconds. Default: %(default)r"
    )
    group.addoption(
        "--benchmark-max-time",
        metavar="SECONDS", type=parse_seconds, default="1.0",
        help="Maximum run time per test - it will be repeated until this total time is reached. It may be "
             "exceeded if test function is very slow or --benchmark-min-rounds is large (it takes precedence). "
             "Default: %(default)r"
    )
    group.addoption(
        "--benchmark-min-rounds",
        metavar="NUM", type=parse_rounds, default=5,
        help="Minimum rounds, even if total time would exceed `--max-time`. Default: %(default)r"
    )
    group.addoption(
        "--benchmark-sort",
        metavar="COL", type=parse_sort, default="min",
        help="Column to sort on. Can be one of: 'min', 'max', 'mean' or 'stddev'. Default: %(default)r"
    )
    group.addoption(
        "--benchmark-group-by",
        metavar="LABEL", default="group",
        help="How to group tests. Can be one of: 'group', 'name', 'fullname', 'func', 'fullfunc' or 'param'."
             " Default: %(default)r"
    )
    group.addoption(
        "--benchmark-timer",
        metavar="FUNC", type=parse_timer, default=str(NameWrapper(default_timer)),
        help="Timer to use when measuring time. Default: %(default)r"
    )
    group.addoption(
        "--benchmark-calibration-precision",
        metavar="NUM", type=int, default=10,
        help="Precision to use when calibrating number of iterations. Precision of 10 will make the timer look 10 times"
             " more accurate, at a cost of less precise measure of deviations. Default: %(default)r"
    )
    group.addoption(
        "--benchmark-warmup",
        metavar="KIND", nargs="?", default=parse_warmup("auto"), type=parse_warmup,
        help="Activates warmup. Will run the test function up to number of times in the calibration phase. "
             "See `--benchmark-warmup-iterations`. Note: Even the warmup phase obeys --benchmark-max-time. "
             "Available KIND: 'auto', 'off', 'on'. Default: 'auto' (automatically activate on PyPy)."
    )
    group.addoption(
        "--benchmark-warmup-iterations",
        metavar="NUM", type=int, default=100000,
        help="Max number of iterations to run in the warmup phase. Default: %(default)r"
    )
    group.addoption(
        "--benchmark-verbose",
        action="store_true", default=False,
        help="Dump diagnostic and progress information."
    )
    group.addoption(
        "--benchmark-disable-gc",
        action="store_true", default=False,
        help="Disable GC during benchmarks."
    )
    group.addoption(
        "--benchmark-skip",
        action="store_true", default=False,
        help="Skip running any tests that contain benchmarks."
    )
    group.addoption(
        "--benchmark-disable",
        action="store_true", default=False,
        help="Disable benchmarks. Benchmarked functions are only ran once and no stats are reported. Use this is you "
             "want to run the test but don't do any benchmarking."
    )
    group.addoption(
        "--benchmark-only",
        action="store_true", default=False,
        help="Only run benchmarks."
    )
    group.addoption(
        "--benchmark-save",
        metavar="NAME", type=parse_save,
        help="Save the current run into 'STORAGE-PATH/counter_NAME.json'."
    )
    tag = get_tag()
    group.addoption(
        "--benchmark-autosave",
        action='store_const', const=tag,
        help="Autosave the current run into 'STORAGE-PATH/counter_%s.json" % tag,
    )
    group.addoption(
        "--benchmark-save-data",
        action="store_true",
        help="Use this to make --benchmark-save and --benchmark-autosave include all the timing data,"
             " not just the stats.",
    )
    group.addoption(
        "--benchmark-compare",
        metavar="NUM", nargs="?", default=[], const=True,
        help="Compare the current run against run NUM or the latest saved run if unspecified."
    )
    group.addoption(
        "--benchmark-compare-fail",
        metavar="EXPR", nargs="+", type=parse_compare_fail,
        help="Fail test if performance regresses according to given EXPR"
             " (eg: min:5%% or mean:0.001 for number of seconds). Can be used multiple times."
    )
    group.addoption(
        "--benchmark-storage",
        metavar="STORAGE-PATH", default="./.benchmarks/%s-%s-%s-%s" % (
            platform.system(),
            platform.python_implementation(),
            ".".join(platform.python_version_tuple()[:2]),
            platform.architecture()[0]
        ),
        help="Specify a different path to store the runs (when --benchmark-save or --benchmark-autosave are used). "
             "Default: %(default)r",
    )
    prefix = "benchmark_%s" % get_current_time()
    group.addoption(
        "--benchmark-histogram",
        action='append', metavar="FILENAME-PREFIX", nargs="?", default=[], const=prefix,
        help="Plot graphs of min/max/avg/stddev over time in FILENAME-PREFIX-test_name.svg. If FILENAME-PREFIX contains"
             " slashes ('/') then directories will be created. Default: %r" % prefix
    )
    group.addoption(
        "--benchmark-json",
        metavar="PATH", type=argparse.FileType('wb'),
        help="Dump a JSON report into PATH. "
             "Note that this will include the complete data (all the timings, not just the stats)."
    )


def pytest_addhooks(pluginmanager):
    from . import hookspec

    method = getattr(pluginmanager, "add_hookspecs", None)
    if method is None:
        method = pluginmanager.addhooks
    method(hookspec)


class BenchmarkStats(object):
    def __init__(self, fixture, iterations, options):
        self.name = fixture.name
        self.fullname = fixture.fullname
        self.group = fixture.group
        self.param = fixture.param

        self.iterations = iterations
        self.stats = Stats()
        self.options = options
        self.fixture = fixture

    def __bool__(self):
        return bool(self.stats)

    def __nonzero__(self):
        return bool(self.stats)

    def get(self, key, default=None):
        try:
            return getattr(self.stats, key)
        except AttributeError:
            return getattr(self, key, default)

    def __getitem__(self, key):
        try:
            return getattr(self.stats, key)
        except AttributeError:
            return getattr(self, key)

    @property
    def has_error(self):
        return self.fixture.has_error

    def json(self, include_data=True):
        if include_data:
            return dict(self.stats.as_dict, data=self.stats.data)
        else:
            return self.stats.as_dict

    def update(self, duration):
        self.stats.update(duration / self.iterations)


class BenchmarkFixture(object):
    _precisions = {}

    @classmethod
    def _get_precision(cls, timer):
        if timer in cls._precisions:
            return cls._precisions[timer]
        else:
            return cls._precisions.setdefault(timer, compute_timer_precision(timer))

    def __init__(self, node, disable_gc, timer, min_rounds, min_time, max_time, warmup, warmup_iterations,
                 calibration_precision, add_stats, logger, warner, disable, group=None):
        self.name = node.name
        self.fullname = node._nodeid
        self.disable = disable
        self.param = node.callspec.id if hasattr(node, 'callspec') else None
        self.group = group
        self.has_error = False

        self._disable_gc = disable_gc
        self._timer = timer.target
        self._min_rounds = min_rounds
        self._max_time = float(max_time)
        self._min_time = float(min_time)
        self._add_stats = add_stats
        self._calibration_precision = calibration_precision
        self._warmup = warmup and warmup_iterations
        self._logger = logger
        self._warner = warner
        self._cleanup_callbacks = []
        self._mode = None

    def _make_runner(self, function_to_benchmark, args, kwargs):
        def runner(loops_range, timer=self._timer):
            gc_enabled = gc.isenabled()
            if self._disable_gc:
                gc.disable()
            tracer = sys.gettrace()
            sys.settrace(None)
            try:
                if loops_range:
                    start = timer()
                    for _ in loops_range:
                        function_to_benchmark(*args, **kwargs)
                    end = timer()
                    return end - start
                else:
                    start = timer()
                    result = function_to_benchmark(*args, **kwargs)
                    end = timer()
                    return end - start, result
            finally:
                sys.settrace(tracer)
                if gc_enabled:
                    gc.enable()

        return runner

    def _make_stats(self, iterations):
        stats = BenchmarkStats(self, iterations=iterations, options={
            "disable_gc": self._disable_gc,
            "timer": self._timer,
            "min_rounds": self._min_rounds,
            "max_time": self._max_time,
            "min_time": self._min_time,
            "warmup": self._warmup,
        })
        self._add_stats(stats)
        return stats

    def __call__(self, function_to_benchmark, *args, **kwargs):
        if self._mode:
            self.has_error = True
            raise FixtureAlreadyUsed(
                "Fixture can only be used once. Previously it was used in %s mode." % self._mode)
        try:
            self._mode = 'benchmark(...)'
            return self._raw(function_to_benchmark, *args, **kwargs)
        except Exception:
            self.has_error = True
            raise

    def pedantic(self, target, args=(), kwargs=None, setup=None, rounds=1, warmup_rounds=0, iterations=1):
        if self._mode:
            self.has_error = True
            raise FixtureAlreadyUsed(
                "Fixture can only be used once. Previously it was used in %s mode." % self._mode)
        try:
            self._mode = 'benchmark.pedantic(...)'
            return self._raw_pedantic(target, args=args, kwargs=kwargs, setup=setup, rounds=rounds,
                                      warmup_rounds=warmup_rounds, iterations=iterations)
        except Exception:
            self.has_error = True
            raise

    def _raw(self, function_to_benchmark, *args, **kwargs):
        if not self.disable:
            runner = self._make_runner(function_to_benchmark, args, kwargs)

            duration, iterations, loops_range = self._calibrate_timer(runner)

            # Choose how many time we must repeat the test
            rounds = int(ceil(self._max_time / duration))
            rounds = max(rounds, self._min_rounds)
            rounds = min(rounds, sys.maxsize)

            stats = self._make_stats(iterations)

            self._logger.debug("  Running %s rounds x %s iterations ..." % (rounds, iterations), yellow=True, bold=True)
            run_start = time.time()
            if self._warmup:
                warmup_rounds = min(rounds, max(1, int(self._warmup / iterations)))
                self._logger.debug("  Warmup %s rounds x %s iterations ..." % (warmup_rounds, iterations))
                for _ in XRANGE(warmup_rounds):
                    runner(loops_range)
            for _ in XRANGE(rounds):
                stats.update(runner(loops_range))
            self._logger.debug("  Ran for %ss." % format_time(time.time() - run_start), yellow=True, bold=True)
        return function_to_benchmark(*args, **kwargs)

    def _raw_pedantic(self, target, args=(), kwargs=None, setup=None, rounds=1, warmup_rounds=0, iterations=1):
        if kwargs is None:
            kwargs = {}

        has_args = bool(args or kwargs)

        if not isinstance(iterations, INT) or iterations < 1:
            raise ValueError("Must have positive int for `iterations`.")

        if not isinstance(rounds, INT) or rounds < 1:
            raise ValueError("Must have positive int for `rounds`.")

        if not isinstance(warmup_rounds, INT) or warmup_rounds < 0:
            raise ValueError("Must have positive int for `warmup_rounds`.")

        if iterations > 1 and setup:
            raise ValueError("Can't use more than 1 `iterations` with a `setup` function.")

        def make_arguments(args=args, kwargs=kwargs):
            if setup:
                maybe_args = setup()
                if maybe_args:
                    if has_args:
                        raise TypeError("Can't use `args` or `kwargs` if `setup` returns the arguments.")
                    args, kwargs = maybe_args
            return args, kwargs

        if self.disable:
            args, kwargs = make_arguments()
            return target(*args, **kwargs)

        stats = self._make_stats(iterations)
        loops_range = XRANGE(iterations) if iterations > 1 else None
        for _ in XRANGE(warmup_rounds):
            args, kwargs = make_arguments()

            runner = self._make_runner(target, args, kwargs)
            runner(loops_range)

        for _ in XRANGE(rounds):
            args, kwargs = make_arguments()

            runner = self._make_runner(target, args, kwargs)
            if loops_range:
                duration = runner(loops_range)
            else:
                duration, result = runner(loops_range)
            stats.update(duration)

        if loops_range:
            args, kwargs = make_arguments()
            result = target(*args, **kwargs)
        return result

    def weave(self, target, **kwargs):
        try:
            import aspectlib
        except ImportError as exc:
            raise ImportError(exc.args, "Please install aspectlib or pytest-benchmark[aspect]")

        def aspect(function):
            def wrapper(*args, **kwargs):
                return self(function, *args, **kwargs)

            return wrapper

        self._cleanup_callbacks.append(aspectlib.weave(target, aspect, **kwargs).rollback)

    patch = weave

    def _cleanup(self):
        while self._cleanup_callbacks:
            callback = self._cleanup_callbacks.pop()
            callback()
        if not self._mode:
            self._logger.warn("BENCHMARK-U1", "Benchmark fixture was not used at all in this test!",
                              warner=self._warner, suspend=True)

    def _calibrate_timer(self, runner):
        timer_precision = self._get_precision(self._timer)
        min_time = max(self._min_time, timer_precision * self._calibration_precision)
        min_time_estimate = min_time * 5 / self._calibration_precision
        self._logger.debug("")
        self._logger.debug("  Timer precision: %ss" % format_time(timer_precision), yellow=True, bold=True)
        self._logger.debug("  Calibrating to target round %ss; will estimate when reaching %ss." % (
            format_time(min_time), format_time(min_time_estimate)), yellow=True, bold=True)

        loops = 1
        while True:
            loops_range = XRANGE(loops)
            duration = runner(loops_range)
            if self._warmup:
                warmup_start = time.time()
                warmup_iterations = 0
                warmup_rounds = 0
                while time.time() - warmup_start < self._max_time and warmup_iterations < self._warmup:
                    duration = min(duration, runner(loops_range))
                    warmup_rounds += 1
                    warmup_iterations += loops
                self._logger.debug("    Warmup: %ss (%s x %s iterations)." % (
                    format_time(time.time() - warmup_start),
                    warmup_rounds, loops
                ))

            self._logger.debug("    Measured %s iterations: %ss." % (loops, format_time(duration)), yellow=True)
            if duration >= min_time:
                break

            if duration >= min_time_estimate:
                # coarse estimation of the number of loops
                loops = int(ceil(min_time * loops / duration))
                self._logger.debug("    Estimating %s iterations." % loops, green=True)
                if loops == 1:
                    # If we got a single loop then bail early - nothing to calibrate if the the
                    # test function is 100 times slower than the timer resolution.
                    loops_range = XRANGE(loops)
                    break
            else:
                loops *= 10
        return duration, loops, loops_range


class Logger(object):
    def __init__(self, verbose, config):
        self.verbose = verbose
        self.term = py.io.TerminalWriter(file=sys.stderr)
        self.capman = config.pluginmanager.getplugin("capturemanager")
        self.pytest_warn = config.warn
        try:
            self.pytest_warn_has_fslocation = 'fslocation' in config.warn.func_code.co_varnames
        except AttributeError:
            self.pytest_warn_has_fslocation = False

    def warn(self, code, text, warner=None, suspend=False, fslocation=None):
        if self.verbose:
            if suspend and self.capman:
                self.capman.suspendcapture(in_=True)
            self.term.line("")
            self.term.sep("-", red=True, bold=True)
            self.term.write(" WARNING: ", red=True, bold=True)
            self.term.line(text, red=True)
            self.term.sep("-", red=True, bold=True)
            if suspend and self.capman:
                self.capman.resumecapture()
        if warner is None:
            warner = self.pytest_warn
        if fslocation and self.pytest_warn_has_fslocation:
            warner(code=code, message=text, fslocation=fslocation)
        else:
            warner(code=code, message=text)

    def error(self, text):
        self.term.line("")
        self.term.sep("-", red=True, bold=True)
        self.term.line(text, red=True, bold=True)
        self.term.sep("-", red=True, bold=True)

    def info(self, text, **kwargs):
        if not kwargs or kwargs == {'bold': True}:
            kwargs['purple'] = True
        self.term.line(text, **kwargs)

    def debug(self, text, **kwargs):
        if self.verbose:
            if self.capman:
                self.capman.suspendcapture(in_=True)
            self.info(text, **kwargs)
            if self.capman:
                self.capman.resumecapture()


class BenchmarkSession(object):
    compare_mapping = None

    def __init__(self, config):
        self.verbose = config.getoption("benchmark_verbose")
        self.logger = Logger(self.verbose, config)
        self.config = config
        self.options = dict(
            min_time=SecondsDecimal(config.getoption("benchmark_min_time")),
            min_rounds=config.getoption("benchmark_min_rounds"),
            max_time=SecondsDecimal(config.getoption("benchmark_max_time")),
            timer=load_timer(config.getoption("benchmark_timer")),
            calibration_precision=config.getoption("benchmark_calibration_precision"),
            disable_gc=config.getoption("benchmark_disable_gc"),
            warmup=config.getoption("benchmark_warmup"),
            warmup_iterations=config.getoption("benchmark_warmup_iterations"),
        )
        self.skip = config.getoption("benchmark_skip")
        self.disable = config.getoption("benchmark_disable")

        if config.getoption("dist", "no") != "no" and not self.skip:
            self.logger.warn(
                "BENCHMARK-U2",
                "Benchmarks are automatically disabled because xdist plugin is active."
                "Benchmarks cannot be performed reliably in a parallelized environment.",
                fslocation="::"
            )
            self.disable = True
        if hasattr(config, "slaveinput"):
            self.disable = True
        if not statistics:
            self.logger.warn(
                "BENCHMARK-U3",
                "Benchmarks are automatically disabled because we could not import `statistics`\n\n%s" %
                statistics_error,
                fslocation="::"
            )
            self.disable = True

        self.only = config.getoption("benchmark_only")
        self.sort = config.getoption("benchmark_sort")
        if self.skip and self.only:
            raise pytest.UsageError("Can't have both --benchmark-only and --benchmark-skip options.")
        if self.disable and self.only:
            raise pytest.UsageError(
                "Can't have both --benchmark-only and --benchmark-disable options. Note that --benchmark-disable is "
                "automatically activated if xdist is on or you're missing the statistics dependency.")
        self._benchmarks = []
        self.group_by = config.getoption("benchmark_group_by")
        self.save = config.getoption("benchmark_save")
        self.autosave = config.getoption("benchmark_autosave")
        self.save_data = config.getoption("benchmark_save_data")
        self.json = config.getoption("benchmark_json")
        self.compare = config.getoption("benchmark_compare")
        self.compare_fail = config.getoption("benchmark_compare_fail")
        self.performance_regressions = []
        self.storage = py.path.local(config.getoption("benchmark_storage"))
        self.storage.ensure(dir=1)
        self.histogram = first_or_value(config.getoption("benchmark_histogram"), False)

    @property
    def benchmarks(self):
        return [bench for bench in self._benchmarks if bench]

    @property
    def storage_fslocation(self):
        return self.storage.relto(os.getcwd())

    @cached_property
    def compare_file(self):
        if self.compare:
            files = self.storage.listdir("[0-9][0-9][0-9][0-9]_*.json", sort=True)
            if files:
                if self.compare is True:
                    files.sort()
                    return files[-1]
                else:
                    files = [f for f in files if str(f.basename).startswith(self.compare)]
                    if len(files) == 1:
                        return files[0]

                    if not files:
                        self.logger.warn("BENCHMARK-C1", "Can't compare. No benchmark files matched %r" % self.compare,
                                         fslocation=self.storage_fslocation)
                    elif len(files) > 1:
                        self.logger.warn(
                            "BENCHMARK-C2", "Can't compare. Too many benchmark files matched %r:\n - %s" % (
                                self.compare, '\n - '.join(map(str, files))
                            ),
                            fslocation=self.storage_fslocation)
            else:
                msg = "Can't compare. No benchmark files in %r. " \
                      "Expected files matching [0-9][0-9][0-9][0-9]_*.json." % str(self.storage)
                if self.compare is True:
                    msg += " Can't load the previous benchmark."
                    code = "BENCHMARK-C3"
                else:
                    msg += " Can't match anything to %r." % self.compare
                    code = "BENCHMARK-C4"
                self.logger.warn(code, msg, fslocation=self.storage_fslocation)
                return

    @property
    def next_num(self):
        files = self.storage.listdir("[0-9][0-9][0-9][0-9]_*.json")
        files.sort(reverse=True)
        if not files:
            return "0001"
        for f in files:
            try:
                return "%04i" % (int(str(f.basename).split('_')[0]) + 1)
            except ValueError:
                raise

    def handle_saving(self):
        if self.json:
            output_json = self.config.hook.pytest_benchmark_generate_json(
                config=self.config,
                benchmarks=self.benchmarks,
                include_data=True
            )
            self.config.hook.pytest_benchmark_update_json(
                config=self.config,
                benchmarks=self.benchmarks,
                output_json=output_json
            )
            with self.json as fh:
                fh.write(json.dumps(output_json, ensure_ascii=True, indent=4).encode())
            self.logger.info("Wrote benchmark data in %s" % self.json, purple=True)

        save = self.save or self.autosave
        if save:
            output_json = self.config.hook.pytest_benchmark_generate_json(
                config=self.config,
                benchmarks=self.benchmarks,
                include_data=self.save_data
            )
            self.config.hook.pytest_benchmark_update_json(
                config=self.config,
                benchmarks=self.benchmarks,
                output_json=output_json
            )
            output_file = self.storage.join("%s_%s.json" % (self.next_num, save))
            assert not output_file.exists()

            with output_file.open('wb') as fh:
                fh.write(json.dumps(output_json, ensure_ascii=True, indent=4).encode())
            self.logger.info("Saved benchmark data in %s" % output_file)

    def handle_loading(self):
        if self.compare_file:
            self.compare_name = self.compare_file.basename.split('_')[0]
            with self.compare_file.open('rU') as fh:
                try:
                    compared_benchmark = json.load(fh)
                except Exception as exc:
                    self.logger.warn("BENCHMARK-C5", "Failed to load %s: %s" % (self.compare_file, exc),
                                     fslocation=self.storage_fslocation)
                    return

            machine_info = self.config.hook.pytest_benchmark_generate_machine_info(config=self.config)
            self.config.hook.pytest_benchmark_update_machine_info(config=self.config, machine_info=machine_info)
            self.config.hook.pytest_benchmark_compare_machine_info(config=self.config, benchmarksession=self,
                                                                   machine_info=machine_info,
                                                                   compared_benchmark=compared_benchmark)
            self.compare_mapping = dict((bench['fullname'], bench) for bench in compared_benchmark['benchmarks'])

            self.logger.info("Comparing against benchmark %s:" % self.compare_file.basename, bold=True)
            self.logger.info("| commit info: %s" % format_dict(compared_benchmark['commit_info']))
            self.logger.info("| saved at: %s" % compared_benchmark['datetime'])
            self.logger.info("| saved using pytest-benchmark %s:" % compared_benchmark['version'])

    def display(self, tr):
        if not self.benchmarks:
            return

        tr.ensure_newline()
        self.handle_saving()
        self.handle_loading()
        if self.benchmarks:
            self.display_results_table(tr)
            self.check_regressions()
            self.handle_histogram()

    def check_regressions(self):
        if self.compare_fail and not self.compare_file:
            raise pytest.UsageError("--benchmark-compare-fail requires valid --benchmark-compare.")

        if self.performance_regressions:
            self.logger.error("Performance has regressed:\n%s" % "\n".join(
                "\t%s - %s" % line for line in self.performance_regressions
            ))
            raise PerformanceRegression("Performance has regressed.")

    def handle_histogram(self):
        if self.histogram:
            from .histogram import make_plot

            history = {}
            for bench_file in self.storage.listdir("[0-9][0-9][0-9][0-9]_*.json"):
                with bench_file.open('rU') as fh:
                    fullname = bench_file.purebasename
                    if '_' in fullname:
                        id_, name = fullname.split('_', 1)
                    else:
                        id_, name = fullname, ''
                    data = history[id_] = json.load(fh)
                    data['name'] = name
                    data['mapping'] = dict((bench['fullname'], bench) for bench in data['benchmarks'])

            for bench in self.benchmarks:
                name = bench.fullname
                for c in "\/:*?<>|":
                    name = name.replace(c, '_').replace('__', '_')
                output_file = py.path.local("%s-%s.svg" % (self.histogram, name)).ensure()

                table = list(self.generate_histogram_table(bench, history, sorted(history)))

                plot = make_plot(
                    bench_name=bench.fullname,
                    table=table,
                    compare=self.compare_file,
                    annotations=history,
                    sort=self.sort,
                    current=HISTOGRAM_CURRENT,
                )
                plot.render_to_file(str(output_file))
                self.logger.info("Generated histogram %s" % output_file, bold=True)

    @staticmethod
    def generate_histogram_table(current, history, sequence):
        for name in sequence:
            trial = history[name]
            for bench in trial["benchmarks"]:
                if bench["fullname"] == current.fullname:
                    found = True
                else:
                    found = False

                if found:
                    yield "%s" % name, bench["stats"]
                    break

        yield HISTOGRAM_CURRENT, current.json()

    def apply_compare(self, benchmarks, compare_name, compare_mapping):
        result = []
        for bench in benchmarks:
            if bench.fullname in compare_mapping:
                stats = compare_mapping[bench.fullname]["stats"]
                result.extend([
                    dict(bench.json(include_data=False),
                         name="{0} ({1})".format(bench.name, "NOW"),
                         iterations=bench.iterations),
                    dict(stats, name="{0} ({1})".format(bench.name, compare_name)),
                ])
                if self.compare_fail:
                    for check in self.compare_fail:
                        fail = check.fails(bench, stats)
                        if fail:
                            self.performance_regressions.append((bench.fullname, fail))
            else:
                result.append(bench)
        return result

    def display_results_table(self, tr):
        tr.write_line("")
        tr.rewrite("Computing stats ...", black=True, bold=True)
        groups = self.config.hook.pytest_benchmark_group_stats(
            config=self.config,
            benchmarks=self.benchmarks,
            group_by=self.group_by
        )
        for line, (group, benchmarks) in report_progress(groups, tr, "Computing stats ... group {pos}/{total}"):
            if self.compare_file:
                benchmarks = self.apply_compare(benchmarks, self.compare_name, self.compare_mapping)
            benchmarks = sorted(benchmarks, key=operator.itemgetter(self.sort))

            worst = {}
            best = {}
            solo = len(benchmarks) == 1
            for line, prop in report_progress(("min", "max", "mean", "median", "iqr", "stddev"), tr, "{line}: {value}", line=line):
                worst[prop] = max(bench[prop] for _, bench in report_progress(
                    benchmarks, tr, "{line} ({pos}/{total})", line=line))
                best[prop] = min(bench[prop] for _, bench in report_progress(
                    benchmarks, tr, "{line} ({pos}/{total})", line=line))
            for line, prop in report_progress(("outliers", "rounds", "iterations"), tr, "{line}: {value}", line=line):
                worst[prop] = max(benchmark[prop] for _, benchmark in report_progress(
                    benchmarks, tr, "{line} ({pos}/{total})", line=line))

            unit, adjustment = time_unit(best.get(self.sort, benchmarks[0][self.sort]))
            labels = {
                "name": "Name (time in %ss)" % unit,
                "min": "Min",
                "max": "Max",
                "mean": "Mean",
                "stddev": "StdDev",
                "rounds": "Rounds",
                "iterations": "Iterations",
                "iqr": "IQR",
                "median": "Median",
                "outliers": "Outliers(*)",
            }
            widths = {
                "name": 3 + max(len(labels["name"]), max(len(benchmark["name"]) for benchmark in benchmarks)),
                "rounds": 2 + max(len(labels["rounds"]), len(str(worst["rounds"]))),
                "iterations": 2 + max(len(labels["iterations"]), len(str(worst["iterations"]))),
                "outliers": 2 + max(len(labels["outliers"]), len(str(worst["outliers"]))),
            }
            for prop in "min", "max", "mean", "stddev", "median", "iqr":
                widths[prop] = 2 + max(len(labels[prop]), max(
                    len(NUMBER_FMT.format(bench[prop] * adjustment))
                    for bench in benchmarks
                ))

            rpadding = 0 if solo else 10
            labels_line = labels["name"].ljust(widths["name"]) + "".join(
                labels[prop].rjust(widths[prop]) + (
                    " " * rpadding
                    if prop not in ["outliers", "rounds", "iterations"]
                    else ""
                )
                for prop in ("min", "max", "mean", "stddev", "median", "iqr", "outliers", "rounds", "iterations")
            )
            tr.rewrite("")
            tr.write_line(
                (" benchmark%(name)s: %(count)s tests " % dict(
                    count=len(benchmarks),
                    name="" if group is None else " %r" % group,
                )).center(len(labels_line), "-"),
                yellow=True,
            )
            tr.write_line(labels_line)
            tr.write_line("-" * len(labels_line), yellow=True)

            for bench in benchmarks:
                has_error = bench.get("has_error")
                tr.write(bench["name"].ljust(widths["name"]), red=has_error, invert=has_error)
                for prop in "min", "max", "mean", "stddev", "median", "iqr":
                    tr.write(
                        ALIGNED_NUMBER_FMT.format(
                            bench[prop] * adjustment,
                            widths[prop],
                            self.compute_baseline_scale(best[prop], bench[prop], rpadding),
                            rpadding
                        ),
                        green=not solo and bench[prop] == best.get(prop),
                        red=not solo and bench[prop] == worst.get(prop),
                        bold=True,
                    )
                for prop in "outliers", "rounds", "iterations":
                    tr.write("{0:>{1}}".format(bench[prop], widths[prop]))
                tr.write("\n")
            tr.write_line("-" * len(labels_line), yellow=True)
            tr.write_line("")
        tr.write_line("(*) Outliers: 1 Standard Deviation from Mean; "
                      "1.5 IQR (InterQuartile Range) from 1st Quartile and 3rd Quartile.", bold=True, black=True)

    def compute_baseline_scale(self, baseline, value, width):
        if not width:
            return ""
        if value == baseline:
            return " (1.0)".ljust(width)

        scale = abs(value / baseline) if baseline else float("inf")
        if scale > 1000:
            if isinf(scale):
                return " (inf)".ljust(width)
            else:
                return " (>1000.0)".ljust(width)
        else:
            return " ({0:.2f})".format(scale).ljust(width)


def pytest_benchmark_compare_machine_info(config, benchmarksession, machine_info, compared_benchmark):
    if compared_benchmark["machine_info"] != machine_info:
        benchmarksession.logger.warn(
            "BENCHMARK-C6",
            "Benchmark machine_info is different. Current: %s VS saved: %s." % (
                format_dict(machine_info),
                format_dict(compared_benchmark["machine_info"]),
            ),
            fslocation=benchmarksession.storage_fslocation
        )

if hasattr(pytest, 'hookimpl'):
    _hookwrapper = pytest.hookimpl(hookwrapper=True)
else:
    _hookwrapper = pytest.mark.hookwrapper


@_hookwrapper
def pytest_runtest_call(item):
    bs = item.config._benchmarksession
    fixure = hasattr(item, "funcargs") and item.funcargs.get("benchmark")
    if isinstance(fixure, BenchmarkFixture):
        if bs.skip:
            pytest.skip("Skipping benchmark (--benchmark-skip active).")
        else:
            yield
    else:
        if bs.only:
            pytest.skip("Skipping non-benchmark (--benchmark-only active).")
        else:
            yield


def pytest_benchmark_group_stats(config, benchmarks, group_by):
    groups = defaultdict(list)
    for bench in benchmarks:
        if group_by == "group":
            groups[bench.group].append(bench)
        elif group_by == "name":
            groups[bench.name].append(bench)
        elif group_by == "func":
            groups[bench.name.split("[")[0]].append(bench)
        elif group_by == "fullfunc":
            groups[bench.fullname.split("[")[0]].append(bench)
        elif group_by == "fullname":
            groups[bench.fullname].append(bench)
        elif group_by == "param":
            groups[bench.param].append(bench)
        else:
            raise NotImplementedError("Unsupported grouping %r." % group_by)
    for grouped_benchmarks in groups.values():
        grouped_benchmarks.sort(key=operator.attrgetter("fullname" if "full" in group_by else "name"))
    return sorted(groups.items(), key=lambda pair: pair[0] or "")


def pytest_terminal_summary(terminalreporter):
    try:
        terminalreporter.config._benchmarksession.display(terminalreporter)
    except PerformanceRegression:
        raise
    except Exception:
        terminalreporter.config._benchmarksession.logger.error("\n%s" % traceback.format_exc())
        raise


def pytest_benchmark_generate_machine_info():
    return {
        "node": platform.node(),
        "processor": platform.processor(),
        "machine": platform.machine(),
        "python_compiler": platform.python_compiler(),
        "python_implementation": platform.python_implementation(),
        "python_version": platform.python_version(),
        "release": platform.release(),
        "system": platform.system()
    }


def pytest_benchmark_generate_commit_info(config):
    return get_commit_info()


def pytest_benchmark_generate_json(config, benchmarks, include_data):
    machine_info = config.hook.pytest_benchmark_generate_machine_info(config=config)
    config.hook.pytest_benchmark_update_machine_info(config=config, machine_info=machine_info)

    commit_info = config.hook.pytest_benchmark_generate_commit_info(config=config)
    config.hook.pytest_benchmark_update_commit_info(config=config, commit_info=commit_info)

    benchmarks_json = []
    output_json = {
        "machine_info": machine_info,
        "commit_info": commit_info,
        "benchmarks": benchmarks_json,
        "datetime": datetime.utcnow().isoformat(),
        "version": __version__,
    }
    for bench in benchmarks:
        if not bench.has_error:
            benchmarks_json.append({
                "group": bench.group,
                "name": bench.name,
                "fullname": bench.fullname,
                "stats": dict(bench.json(include_data=include_data), iterations=bench.iterations),
                "options": dict(
                    (k, v.__name__ if callable(v) else v) for k, v in bench.options.items()
                )
            })
    return output_json


@pytest.fixture(scope="function")
def benchmark(request):
    bs = request.config._benchmarksession

    if bs.skip:
        pytest.skip("Benchmarks are skipped (--benchmark-skip was used).")
    else:
        node = request.node
        marker = node.get_marker("benchmark")
        options = marker.kwargs if marker else {}
        if "timer" in options:
            options["timer"] = NameWrapper(options["timer"])
        fixture = BenchmarkFixture(
            node,
            add_stats=bs._benchmarks.append,
            logger=bs.logger,
            warner=request.node.warn,
            disable=bs.disable,
            **dict(bs.options, **options)
        )
        request.addfinalizer(fixture._cleanup)
        return fixture


@pytest.fixture(scope="function")
def benchmark_weave(benchmark):
    return benchmark.weave


def pytest_runtest_setup(item):
    marker = item.get_marker("benchmark")
    if marker:
        if marker.args:
            raise ValueError("benchmark mark can't have positional arguments.")
        for name in marker.kwargs:
            if name not in (
                    "max_time", "min_rounds", "min_time", "timer", "group", "disable_gc", "warmup",
                    "warmup_iterations", "calibration_precision"):
                raise ValueError("benchmark mark can't have %r keyword argument." % name)


@pytest.mark.trylast  # force the other plugins to initialise, fixes issue with capture not being properly initialised
def pytest_configure(config):
    config.addinivalue_line("markers", "benchmark: mark a test with custom benchmark settings.")
    config._benchmarksession = BenchmarkSession(config)
    config.pluginmanager.register(config._benchmarksession, "pytest-benchmark")