python sdk performance benchmarks

benchmark.py

import random
import string
import uuid
import json
import time
import os

import cProfile
import pstats

import memcache
from couchbase.couchbaseclient import CouchbaseClient
from threading import Thread
from multiprocessing import Process
from pylibcb import Client

# Benchmark parameters
HOST = "127.0.0.1"
ITEMS = 50000  # Total number of documents to create/read
VALUES_PER_DOC = 1
VALUE_LENGTH = 92  # together with previous parameter it defines document size

def prepare_data():
    # Test data
    print "Preparing data..."

    # Random string generator
    randstr = lambda length: ''.join(random.choice(string.letters + string.digits)
                                     for _ in range(length))

    data = dict()
    for item in range(ITEMS):
        key = uuid.uuid4().hex
        value = dict((randstr(10), randstr(VALUE_LENGTH))
                     for _ in range(VALUES_PER_DOC))
        data[key] = json.dumps(value)

    return data

def do_sets(cb, data):
    # Insert all items
    print "Loading data..."

    start_time = time.time()
    for key, value in data.items():
        #cb.set(key, 0, 0, value)
        cb.set(key, value)
    end_time = time.time()
    print "Set cmds/sec: {0}".format(ITEMS/(end_time - start_time))

def do_gets(cb, data):
    # Read all items from database
    print "Reading data..."

    start_time = time.time()
    for key in data:
        cb.get(key)
    end_time = time.time()

    print "Get cmds/sec: {0}".format(ITEMS/(end_time - start_time))

def main():
    # Establish connection
    cb = CouchbaseClient("http://{0}:8091/pools/default".format(HOST),
                         "default", "password", True)

    #cb = memcache.Client(["{0}:11211".format(HOST)], debug=0)
    cb = Client(host=HOST, user='Administrator', passwd='password',
                bucket='default')

    dataset = [prepare_data() for _ in range(3)]
    threads = list()

    for data in dataset:
        t = Thread(target=do_sets, args=(cb, data))
        threads.append(t)
        t.start()

    for thread in threads:
        thread.join()

    """
    data = prepare_data()
    
    do_sets(cb, data)"""

    #cb.__del__()

    #cProfile.runctx('do_sets(cb, data)', globals(), locals(), 'cb_profile')

if __name__ == "__main__":
    main()

    if False:
        p = pstats.Stats('cb_profile')
        p.sort_stats('cumulative').print_stats(10)
        p.sort_stats('time').print_stats(10)
        p.sort_stats('calls').print_stats(10)

    os._exit(0)