diff options
Diffstat (limited to 'scripts/buildstats-diff')
| -rwxr-xr-x | scripts/buildstats-diff | 342 |
1 files changed, 54 insertions, 288 deletions
diff --git a/scripts/buildstats-diff b/scripts/buildstats-diff index adeba44988..2f6498ab67 100755 --- a/scripts/buildstats-diff +++ b/scripts/buildstats-diff @@ -4,26 +4,23 @@ # # Copyright (c) 2016, Intel Corporation. # -# This program is free software; you can redistribute it and/or modify it -# under the terms and conditions of the GNU General Public License, -# version 2, as published by the Free Software Foundation. -# -# This program is distributed in the hope it will be useful, but WITHOUT -# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or -# FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for -# more details. +# SPDX-License-Identifier: GPL-2.0-only # + import argparse import glob -import json import logging import math import os -import re import sys -from collections import namedtuple from operator import attrgetter +# Import oe libs +scripts_path = os.path.dirname(os.path.realpath(__file__)) +sys.path.append(os.path.join(scripts_path, 'lib')) +from buildstats import BuildStats, diff_buildstats, taskdiff_fields, BSVerDiff + + # Setup logging logging.basicConfig(level=logging.INFO, format="%(levelname)s: %(message)s") log = logging.getLogger() @@ -34,180 +31,16 @@ class ScriptError(Exception): pass -taskdiff_fields = ('pkg', 'pkg_op', 'task', 'task_op', 'value1', 'value2', - 'absdiff', 'reldiff') -TaskDiff = namedtuple('TaskDiff', ' '.join(taskdiff_fields)) - - -class BSTask(dict): - def __init__(self, *args, **kwargs): - self['start_time'] = None - self['elapsed_time'] = None - self['status'] = None - self['iostat'] = {} - self['rusage'] = {} - self['child_rusage'] = {} - super(BSTask, self).__init__(*args, **kwargs) - - @property - def cputime(self): - """Sum of user and system time taken by the task""" - return self['rusage']['ru_stime'] + self['rusage']['ru_utime'] + \ - self['child_rusage']['ru_stime'] + self['child_rusage']['ru_utime'] - - @property - def walltime(self): - """Elapsed wall clock time""" - return self['elapsed_time'] - - @property - def read_bytes(self): - """Bytes read from the block layer""" - return self['iostat']['read_bytes'] - - @property - def write_bytes(self): - """Bytes written to the block layer""" - return self['iostat']['write_bytes'] - - @property - def read_ops(self): - """Number of read operations on the block layer""" - return self['rusage']['ru_inblock'] + self['child_rusage']['ru_inblock'] - - @property - def write_ops(self): - """Number of write operations on the block layer""" - return self['rusage']['ru_oublock'] + self['child_rusage']['ru_oublock'] - - -def read_buildstats_file(buildstat_file): - """Convert buildstat text file into dict/json""" - bs_task = BSTask() - log.debug("Reading task buildstats from %s", buildstat_file) - with open(buildstat_file) as fobj: - for line in fobj.readlines(): - key, val = line.split(':', 1) - val = val.strip() - if key == 'Started': - start_time = float(val) - bs_task['start_time'] = start_time - elif key == 'Ended': - end_time = float(val) - elif key.startswith('IO '): - split = key.split() - bs_task['iostat'][split[1]] = int(val) - elif key.find('rusage') >= 0: - split = key.split() - ru_key = split[-1] - if ru_key in ('ru_stime', 'ru_utime'): - val = float(val) - else: - val = int(val) - ru_type = 'rusage' if split[0] == 'rusage' else \ - 'child_rusage' - bs_task[ru_type][ru_key] = val - elif key == 'Status': - bs_task['status'] = val - bs_task['elapsed_time'] = end_time - start_time - return bs_task - - -def read_buildstats_dir(bs_dir): - """Read buildstats directory""" - def split_nevr(nevr): - """Split name and version information from recipe "nevr" string""" - n_e_v, revision = nevr.rsplit('-', 1) - match = re.match(r'^(?P<name>\S+)-((?P<epoch>[0-9]{1,5})_)?(?P<version>[0-9]\S*)$', - n_e_v) - if not match: - # If we're not able to parse a version starting with a number, just - # take the part after last dash - match = re.match(r'^(?P<name>\S+)-((?P<epoch>[0-9]{1,5})_)?(?P<version>[^-]+)$', - n_e_v) - name = match.group('name') - version = match.group('version') - epoch = match.group('epoch') - return name, epoch, version, revision - - if not os.path.isfile(os.path.join(bs_dir, 'build_stats')): - raise ScriptError("{} does not look like a buildstats directory".format(bs_dir)) - - log.debug("Reading buildstats directory %s", bs_dir) - - buildstats = {} - subdirs = os.listdir(bs_dir) - for dirname in subdirs: - recipe_dir = os.path.join(bs_dir, dirname) - if not os.path.isdir(recipe_dir): - continue - name, epoch, version, revision = split_nevr(dirname) - recipe_bs = {'nevr': dirname, - 'name': name, - 'epoch': epoch, - 'version': version, - 'revision': revision, - 'tasks': {}} - for task in os.listdir(recipe_dir): - recipe_bs['tasks'][task] = [read_buildstats_file( - os.path.join(recipe_dir, task))] - if name in buildstats: - raise ScriptError("Cannot handle multiple versions of the same " - "package ({})".format(name)) - buildstats[name] = recipe_bs - - return buildstats - - -def bs_append(dst, src): - """Append data from another buildstats""" - if set(dst.keys()) != set(src.keys()): - raise ScriptError("Refusing to join buildstats, set of packages is " - "different") - for pkg, data in dst.items(): - if data['nevr'] != src[pkg]['nevr']: - raise ScriptError("Refusing to join buildstats, package version " - "differs: {} vs. {}".format(data['nevr'], src[pkg]['nevr'])) - if set(data['tasks'].keys()) != set(src[pkg]['tasks'].keys()): - raise ScriptError("Refusing to join buildstats, set of tasks " - "in {} differ".format(pkg)) - for taskname, taskdata in data['tasks'].items(): - taskdata.extend(src[pkg]['tasks'][taskname]) - - -def read_buildstats_json(path): - """Read buildstats from JSON file""" - buildstats = {} - with open(path) as fobj: - bs_json = json.load(fobj) - for recipe_bs in bs_json: - if recipe_bs['name'] in buildstats: - raise ScriptError("Cannot handle multiple versions of the same " - "package ({})".format(recipe_bs['name'])) - - if recipe_bs['epoch'] is None: - recipe_bs['nevr'] = "{}-{}-{}".format(recipe_bs['name'], recipe_bs['version'], recipe_bs['revision']) - else: - recipe_bs['nevr'] = "{}-{}_{}-{}".format(recipe_bs['name'], recipe_bs['epoch'], recipe_bs['version'], recipe_bs['revision']) - - for task, data in recipe_bs['tasks'].copy().items(): - recipe_bs['tasks'][task] = [BSTask(data)] - - buildstats[recipe_bs['name']] = recipe_bs - - return buildstats - - def read_buildstats(path, multi): """Read buildstats""" if not os.path.exists(path): raise ScriptError("No such file or directory: {}".format(path)) if os.path.isfile(path): - return read_buildstats_json(path) + return BuildStats.from_file_json(path) if os.path.isfile(os.path.join(path, 'build_stats')): - return read_buildstats_dir(path) + return BuildStats.from_dir(path) # Handle a non-buildstat directory subpaths = sorted(glob.glob(path + '/*')) @@ -222,91 +55,66 @@ def read_buildstats(path, multi): bs = None for subpath in subpaths: if os.path.isfile(subpath): - tmpbs = read_buildstats_json(subpath) + _bs = BuildStats.from_file_json(subpath) else: - tmpbs = read_buildstats_dir(subpath) - if not bs: - bs = tmpbs + _bs = BuildStats.from_dir(subpath) + if bs is None: + bs = _bs else: - log.debug("Joining buildstats") - bs_append(bs, tmpbs) - + bs.aggregate(_bs) if not bs: raise ScriptError("No buildstats found under {}".format(path)) + return bs def print_ver_diff(bs1, bs2): """Print package version differences""" - pkgs1 = set(bs1.keys()) - pkgs2 = set(bs2.keys()) - new_pkgs = pkgs2 - pkgs1 - deleted_pkgs = pkgs1 - pkgs2 - - echanged = [] - vchanged = [] - rchanged = [] - unchanged = [] - common_pkgs = pkgs2.intersection(pkgs1) - if common_pkgs: - for pkg in common_pkgs: - if bs1[pkg]['epoch'] != bs2[pkg]['epoch']: - echanged.append(pkg) - elif bs1[pkg]['version'] != bs2[pkg]['version']: - vchanged.append(pkg) - elif bs1[pkg]['revision'] != bs2[pkg]['revision']: - rchanged.append(pkg) - else: - unchanged.append(pkg) - maxlen = max([len(pkg) for pkg in pkgs1.union(pkgs2)]) + diff = BSVerDiff(bs1, bs2) + + maxlen = max([len(r) for r in set(bs1.keys()).union(set(bs2.keys()))]) fmt_str = " {:{maxlen}} ({})" -# if unchanged: -# print("\nUNCHANGED PACKAGES:") -# print("-------------------") -# maxlen = max([len(pkg) for pkg in unchanged]) -# for pkg in sorted(unchanged): -# print(fmt_str.format(pkg, bs2[pkg]['nevr'], maxlen=maxlen)) - - if new_pkgs: - print("\nNEW PACKAGES:") - print("-------------") - for pkg in sorted(new_pkgs): - print(fmt_str.format(pkg, bs2[pkg]['nevr'], maxlen=maxlen)) - - if deleted_pkgs: - print("\nDELETED PACKAGES:") - print("-----------------") - for pkg in sorted(deleted_pkgs): - print(fmt_str.format(pkg, bs1[pkg]['nevr'], maxlen=maxlen)) + + if diff.new: + print("\nNEW RECIPES:") + print("------------") + for name, val in sorted(diff.new.items()): + print(fmt_str.format(name, val.nevr, maxlen=maxlen)) + + if diff.dropped: + print("\nDROPPED RECIPES:") + print("----------------") + for name, val in sorted(diff.dropped.items()): + print(fmt_str.format(name, val.nevr, maxlen=maxlen)) fmt_str = " {0:{maxlen}} {1:<20} ({2})" - if rchanged: + if diff.rchanged: print("\nREVISION CHANGED:") print("-----------------") - for pkg in sorted(rchanged): - field1 = "{} -> {}".format(pkg, bs1[pkg]['revision'], bs2[pkg]['revision']) - field2 = "{} -> {}".format(bs1[pkg]['nevr'], bs2[pkg]['nevr']) - print(fmt_str.format(pkg, field1, field2, maxlen=maxlen)) + for name, val in sorted(diff.rchanged.items()): + field1 = "{} -> {}".format(val.left.revision, val.right.revision) + field2 = "{} -> {}".format(val.left.nevr, val.right.nevr) + print(fmt_str.format(name, field1, field2, maxlen=maxlen)) - if vchanged: + if diff.vchanged: print("\nVERSION CHANGED:") print("----------------") - for pkg in sorted(vchanged): - field1 = "{} -> {}".format(bs1[pkg]['version'], bs2[pkg]['version']) - field2 = "{} -> {}".format(bs1[pkg]['nevr'], bs2[pkg]['nevr']) - print(fmt_str.format(pkg, field1, field2, maxlen=maxlen)) + for name, val in sorted(diff.vchanged.items()): + field1 = "{} -> {}".format(val.left.version, val.right.version) + field2 = "{} -> {}".format(val.left.nevr, val.right.nevr) + print(fmt_str.format(name, field1, field2, maxlen=maxlen)) - if echanged: + if diff.echanged: print("\nEPOCH CHANGED:") print("--------------") - for pkg in sorted(echanged): - field1 = "{} -> {}".format(bs1[pkg]['epoch'], bs2[pkg]['epoch']) - field2 = "{} -> {}".format(bs1[pkg]['nevr'], bs2[pkg]['nevr']) - print(fmt_str.format(pkg, field1, field2, maxlen=maxlen)) + for name, val in sorted(diff.echanged.items()): + field1 = "{} -> {}".format(val.left.epoch, val.right.epoch) + field2 = "{} -> {}".format(val.left.nevr, val.right.nevr) + print(fmt_str.format(name, field1, field2, maxlen=maxlen)) -def print_task_diff(bs1, bs2, val_type, min_val=0, min_absdiff=0, sort_by=('absdiff',)): +def print_task_diff(bs1, bs2, val_type, min_val=0, min_absdiff=0, sort_by=('absdiff',), only_tasks=[]): """Diff task execution times""" def val_to_str(val, human_readable=False): """Convert raw value to printable string""" @@ -343,12 +151,11 @@ def print_task_diff(bs1, bs2, val_type, min_val=0, min_absdiff=0, sort_by=('absd """Get cumulative sum of all tasks""" total = 0.0 for recipe_data in buildstats.values(): - for bs_task in recipe_data['tasks'].values(): - total += sum([getattr(b, val_type) for b in bs_task]) / len(bs_task) + for name, bs_task in recipe_data.tasks.items(): + if not only_tasks or name in only_tasks: + total += getattr(bs_task, val_type) return total - tasks_diff = [] - if min_val: print("Ignoring tasks less than {} ({})".format( val_to_str(min_val, True), val_to_str(min_val))) @@ -357,49 +164,7 @@ def print_task_diff(bs1, bs2, val_type, min_val=0, min_absdiff=0, sort_by=('absd val_to_str(min_absdiff, True), val_to_str(min_absdiff))) # Prepare the data - pkgs = set(bs1.keys()).union(set(bs2.keys())) - for pkg in pkgs: - tasks1 = bs1[pkg]['tasks'] if pkg in bs1 else {} - tasks2 = bs2[pkg]['tasks'] if pkg in bs2 else {} - if not tasks1: - pkg_op = '+ ' - elif not tasks2: - pkg_op = '- ' - else: - pkg_op = ' ' - - for task in set(tasks1.keys()).union(set(tasks2.keys())): - task_op = ' ' - if task in tasks1: - # Average over all values - val1 = [getattr(b, val_type) for b in bs1[pkg]['tasks'][task]] - val1 = sum(val1) / len(val1) - else: - task_op = '+ ' - val1 = 0 - if task in tasks2: - # Average over all values - val2 = [getattr(b, val_type) for b in bs2[pkg]['tasks'][task]] - val2 = sum(val2) / len(val2) - else: - val2 = 0 - task_op = '- ' - - if val1 == 0: - reldiff = float('inf') - else: - reldiff = 100 * (val2 - val1) / val1 - - if max(val1, val2) < min_val: - log.debug("Filtering out %s:%s (%s)", pkg, task, - val_to_str(max(val1, val2))) - continue - if abs(val2 - val1) < min_absdiff: - log.debug("Filtering out %s:%s (difference of %s)", pkg, task, - val_to_str(val2-val1)) - continue - tasks_diff.append(TaskDiff(pkg, pkg_op, task, task_op, val1, val2, - val2-val1, reldiff)) + tasks_diff = diff_buildstats(bs1, bs2, val_type, min_val, min_absdiff, only_tasks) # Sort our list for field in reversed(sort_by): @@ -484,6 +249,8 @@ Script for comparing buildstats of two separate builds.""" parser.add_argument('--multi', action='store_true', help="Read all buildstats from the given paths and " "average over them") + parser.add_argument('--only-task', dest='only_tasks', metavar='TASK', action='append', default=[], + help="Only include TASK in report. May be specified multiple times") parser.add_argument('buildstats1', metavar='BUILDSTATS1', help="'Left' buildstat") parser.add_argument('buildstats2', metavar='BUILDSTATS2', help="'Right' buildstat") @@ -502,7 +269,6 @@ Script for comparing buildstats of two separate builds.""" return args - def main(argv=None): """Script entry point""" args = parse_args(argv) @@ -526,7 +292,7 @@ def main(argv=None): print_ver_diff(bs1, bs2) else: print_task_diff(bs1, bs2, args.diff_attr, args.min_val, - args.min_absdiff, sort_by) + args.min_absdiff, sort_by, args.only_tasks) except ScriptError as err: log.error(str(err)) return 1 |