gpsum

#!/usr/bin/python3 -u

# Copyright 2016 Hewlett-Packard Development Company, L.P.
# Use of this script is subject to HP Terms of Use at
# http://www8.hp.com/us/en/privacy/terms-of-use.html.

# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at

#    http://www.apache.org/licenses/LICENSE-2.0

# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

import re
import sys
from optparse import OptionParser, OptionGroup

def error(text):
    print(text)
    sys.exit()

def main(argv):

    global options

    parser = OptionParser()
    parser.add_option('-c',   dest='col',      help='column to summarize [cilr]', default='r')
    parser.add_option('-f',   dest='file',     help='file containing test data', default='')
    parser.add_option('-v',   dest='vertical', help='vertical output', action='store_true')
    try:
        (options, args) = parser.parse_args(argv)
    except:
        error('invalid option')

    if options.file == '':
        error('-f required')

    try:
        file = open(options.file, 'r')
    except:
        error("Couldn't open %s" % options.file)

    data = {}
    obj_sizes = {}
    num_threads = {}
    tests = {}
    median = offset = 0

    # assumes getput output w/o --ldist
    gpshift = 1

    for line in file:

        if re.match("#gpsuite", line):
            gpshift = 0
            offset = 11

        # ignore blank lines and comments
        line = line[:-1]
        if line == '' or re.match("^#", line):
            continue

        # ignore api errors
        if re.search('API', line):
            continue

        fields = line.split()
        test = fields[0 + gpshift]

        # gotta do test this way because strings could be in gpsuite headers
        if test == 'put' or test == 'get' or test == 'del' or re.match('[pg]\d+', test):
            if test == 'put':
                tname = '1' + test
            elif test == 'get':
                tname = '2' + test
            elif test == 'del':
                tname = '3' + test
            elif re.match('p\d+g\d+', test):
                tname = '4' + test
            elif re.match('p\d+', test):
                tname = '5' + test
            elif re.match('g\d+', test):
                tname = '6' + test

            if tname not in tests:
                tests[tname] = ''

            clients = int(fields[1 + gpshift])
            procs = int(fields[2 + gpshift])
            osize = fields[3 + gpshift].lower()
            match = re.match('(\d+)(\S)', osize)
            if match:
                osize = '%s%07d' % (match.group(2), int(match.group(1)))

            threads = procs
            
            key = '%s-%04d-%s' % (tname, threads, osize)
            if not key in data:
                data[key]={}

            data[key]['rate'] = fields[6 + gpshift]
            data[key]['iops'] = fields[8 + gpshift]
            data[key]['lat'] = fields[10 + gpshift]
            data[key]['range'] = fields[12 + gpshift]
            data[key]['cpu'] = fields[13+gpshift+offset]

            # remember all uniqe object sizes and thread counts so we can sort
            obj_sizes[osize] = ''
            num_threads['%04d' % threads] = '%s-%s' % (clients, procs)

    numcols = len(obj_sizes)
    if options.col == 'c':
        col_name = '%CPU'
        col_type = 'cpu'
    elif options.col == 'i':
        col_name = 'IOPS'
        col_type = 'iops'
    elif options.col == 'l':
        col_name = 'Latency'
        col_type = 'lat'
    elif options.col == 'r':
        col_name = 'MB/sec'
        col_type = 'rate'
    elif options.col == 'R':
        col_name = 'LatRange'
        col_type = 'range'

    if options.vertical:
        for tname in sorted(tests):
            if tname not in tests:
                continue
            test = tname[1:]

            # headers
            print("Test: %s  %s" % (test, col_name))
            print('%4s %3s %3s ' % ('Totl', 'Clt', 'Prc'), end=' ')
            for osize in sorted(obj_sizes):
                size = str(int(osize[1:])) + osize[0]
                print(' %7s' % size, end=' ')
            print()

            # print one line
            for threads in sorted(num_threads):
                clients, procs = num_threads[threads].split('-')
                print('%4d %3s %3d ' % (int(procs), clients, int(procs) / int(clients)), end=' ')

                for osize in sorted(obj_sizes):
                    key = '%s-%04s-%s' % (tname, threads, osize)

                    try:
                        value = data[key][col_type]
                    except KeyError:
                        value = '-'
                    print(' %7s' % value, end=' ')

                print()
            print()

    else:
        # headers
        data_width = 9 * len(obj_sizes)
        line = '*** %-4s ***' % col_name
        for tname in sorted(tests):
            if tname not in tests:
                continue
            test = tname[1:]

            line += ' '
            line += '{type:{width}}'.format(type='    *** %s ***' % test, width=data_width)
        print(line)

        line = '%4s %3s %3s ' % ('Totl', 'Clt', 'Prc')
        for tname in sorted(tests):
            if tname not in tests:
                continue
            test = tname[1:]

            line += ' '
            for osize in sorted(obj_sizes):
                size = str(int(osize[1:])) + osize[0]
                line += ' %8s' % size
        print(line)

        # data
        for threads in sorted(num_threads):
            clients, procs = num_threads[threads].split('-')
            line = '%4d %3s %3d ' % (int(procs), clients, int(procs) / int(clients))

            for tname in sorted(tests):
                if tname not in tests:
                    continue
                test = tname[1:]

                line += ' '
                for osize in sorted(obj_sizes):
                    key = '%s-%04s-%s' % (tname, threads, osize)
                    try:
                        value = data[key][col_type]
                    except KeyError:
                        value = '-'
                    line += ' %8s' % value
            print(line)


if __name__ == "__main__":

    main(sys.argv[1:])