Benchmarking of YAML vs JSON parsing

benchmark-parsers.py

#!/usr/bin/env python
#
# Running of the script against the following files (OONI reports):
# 87M     2015-12-22.json
# 97M     2015-12-22.yaml
#
# vanilla json: 1.37932395935
# ultra json: 0.421966075897
# simple json: 1.23581790924
# yaml: 193.864903927
# cyaml: 4.40925312042

import time
import json
import sys

target_file = sys.argv[1]
target_file_data = []
target_file_data_yaml = []

with open(target_file) as f:
    for line in f:
        target_file_data.append(line.strip())

with open(target_file.replace(".json", ".yaml")) as f:
    entry = ""
    for line in f:
        if line.startswith("---"):
            continue
        if line.startswith("..."):
            target_file_data_yaml.append(entry)
            entry = ""
        else:
            entry += line

def benchmark_vanilla_json():
    start_time = time.time()
    for line in target_file_data:
        json.loads(line)
    return time.time() - start_time

def benchmark_ujson():
    import ujson
    start_time = time.time()
    for line in target_file_data:
        ujson.loads(line)
    return time.time() - start_time

def benchmark_simplejson():
    import simplejson
    start_time = time.time()
    for line in target_file_data:
        simplejson.loads(line)
    return time.time() - start_time

def benchmark_yaml():
    import yaml
    start_time = time.time()
    for line in target_file_data_yaml:
        yaml.load(line)
    return time.time() - start_time

def benchmark_cyaml():
    import yaml
    from yaml import CLoader
    start_time = time.time()
    for line in target_file_data_yaml:
        yaml.load(line, loader=CLoader)
    return time.time() - start_time

print "vanilla json: %s" % benchmark_vanilla_json()
print "ultra json: %s" % benchmark_ujson()
print "simple json: %s" % benchmark_simplejson()
print "yaml: %s" % benchmark_yaml()
print "cyaml: %s" % benchmark_cyaml()