YCSB/bin/ycsb

#!/usr/bin/env python

import os
import sys
import subprocess

BASE_URL = "https://github.com/brianfrankcooper/YCSB/tree/master/"
COMMANDS = {
    "load" : {
        "command"     : "-load",
        "description" : "Execute the load phase",
    },
    "run" : {
        "command"     : "-t",
        "description" : "Execute the transaction phase",
    },
}

DATABASES = {
    "basic"        : "com.yahoo.ycsb.BasicDB",
    "cassandra-7"  : "com.yahoo.ycsb.db.CassandraClient7",
    "cassandra-8"  : "com.yahoo.ycsb.db.CassandraClient8",
    "cassandra-10" : "com.yahoo.ycsb.db.CassandraClient10",
    "gemfire"      : "com.yahoo.ycsb.db.GemFireClient",
    "hbase"        : "com.yahoo.ycsb.db.HBaseClient",
    "infinispan"   : "com.yahoo.ycsb.db.InfinispanClient",
    "jdbc"         : "com.yahoo.ycsb.db.JdbcDBClient",
    "mapkeeper"    : "com.yahoo.ycsb.db.MapKeeperClient",
    "mongodb"      : "com.yahoo.ycsb.db.MongoDbClient",
    "nosqldb"      : "com.yahoo.ycsb.db.NoSqlDbClient",
    "redis"        : "com.yahoo.ycsb.db.RedisClient", 
    "voldemort"    : "com.yahoo.ycsb.db.VoldemortClient", 
}

OPTIONS = {
    "-p key=value" : "Override workload property",
    "-s"           : "Print status to stderr",
    "-target n"    : "Target ops/sec (default: unthrottled)",
    "-threads n"   : "Number of client threads (default: 1)",
}

def usage():
    print "Usage: %s command database workload-file [options]" % sys.argv[0]

    print "\nCommands:"
    for command in sorted(COMMANDS.keys()):
        print "    %s %s" % (command.ljust(13), COMMANDS[command]["description"])

    print "\nDatabases:"
    for db in sorted(DATABASES.keys()):
        print "    {0:13} {1}".format(db, BASE_URL + db.split("-")[0])

    print """\nWorkload Files:
    There are various predefined workloads under workloads/ directory.
    See https://github.com/brianfrankcooper/YCSB/wiki/Core-Properties
    for the list of workload properties."""

    print "\nOptions:"
    for option in sorted(OPTIONS.keys()):
        print "    %s %s" % (option.ljust(13), OPTIONS[option])
    sys.exit(1)

def find_jars(dir, database):
    jars = []
    for (dirpath, dirnames, filenames) in os.walk(dir):
        if dirpath.endswith("conf"):
            jars.append(dirpath)
        for filename in filenames:
            if filename.endswith(".jar") and \
               (filename.startswith("core") or \
                filename.startswith(database.split("-")[0]) or \
                not "binding" in filename):
                jars.append(os.path.join(dirpath, filename))
    return jars

def get_ycsb_home():
    dir = os.path.abspath(os.path.dirname(sys.argv[0]))
    while "CHANGELOG" not in os.listdir(dir):
        dir = os.path.join(dir, os.path.pardir)
    return os.path.abspath(dir)

if len(sys.argv) < 4:
    usage()
if sys.argv[1] not in COMMANDS:
    print "ERROR: Command '%s' not found" % sys.argv[1]
    usage()
if sys.argv[2] not in DATABASES:
    print "ERROR: Database '%s' not found" % sys.argv[2]
    usage()

ycsb_home = get_ycsb_home()
command = COMMANDS[sys.argv[1]]["command"]
database = sys.argv[2]
db_classname = DATABASES[database]
workload = sys.argv[3]
options = sys.argv[4:]

ycsb_command = ["java", "-cp", ":".join(find_jars(ycsb_home, database)), \
                "com.yahoo.ycsb.Client", command, "-db", db_classname, \
                "-P", workload] + options
print " ".join(ycsb_command)
subprocess.call(ycsb_command)