claroaktests.py

from re import findall
from time import time
from datetime import datetime
from csv import writer
from pprint import pformat
from requests import get


### <CONFING>

EDGE_HOSTNAMES = (
    'brss.clarovideo.com.edgesuite.net', # UNINET
    'brss2.clarovideo.com.edgesuite.net', # UNINET
    #'brss1.clarovideo.com.edgesuite.net', # BRAZILIAN HOST
)

URLS = (
    '/multimediav81/plataforma_vod/ISM/WMP4H00984MTDS_full/WMP4H00984MTDS_full.ism/QualityLevels(2524000)/Fragments(video=300300000)',
    '/multimediav81/plataforma_vod/ISM/WMP4H00984MTDS_full/WMP4H00984MTDS_full.ism/QualityLevels(2524000)/Fragments(video=380380000)',
    '/multimediav81/plataforma_vod/ISM/WMP4H00984MTDS_full/WMP4H00984MTDS_full.ism/QualityLevels(2524000)/Fragments(video=440440000)',
    '/multimediav81/plataforma_vod/ISM/WMP4H00984MTDS_full/WMP4H00984MTDS_full.ism/QualityLevels(2524000)/Fragments(video=480480000)',
    '/multimediav81/plataforma_vod/ISM/WMP4H00984MTDS_full/WMP4H00984MTDS_full.ism/QualityLevels(2524000)/Fragments(video=80080000)',
    '/multimediav81/plataforma_vod/ISM/WMP4H00984MTDS_full/WMP4H00984MTDS_full.ism/QualityLevels(2524000)/Fragments(video=180180000)',
    '/multimediav81/plataforma_vod/ISM/WMP4H00984MTDS_full/WMP4H00984MTDS_full.ism/QualityLevels(2524000)/Fragments(video=240240000)',
)

### </CONFIG>

AKAMAI_PRAGMA = (
    'akamai-x-cache-on', 'akamai-x-cache-remote-on',
    'akamai-x-check-cacheable', 'akamai-x-get-cache-key',
    'akamai-x-get-extracted-values', 'akamai-x-get-nonces', 
    'akamai-x-get-ssl-client-session-id', 'akamai-x-get-true-cache-key', 
    'akamai-x-serial-no', 'akamai-x-cache-on', 'akamai-x-cache-remote-on', 
    'akamai-x-check-cacheable', 'akamai-x-get-cache-key', 
    'akamai-x-get-extracted-values', 'akamai-x-get-nonces', 
    'akamai-x-get-ssl-client-session-id', 
    'akamai-x-get-true-cache-key', 'akamai-x-serial-no'
)


def request(url):
    """Returns a tuple of:
    * timestamp
    * url
    * time taken
    * content length in KB
    * cache code, e.g. TCP_HIT
    * edge ip
    * raw headers
    """
    headers = {
        'Host': findall(r'http://([^/]+)\.edgesuite\.net.+', url)[0],
        'Pragma': ','.join(AKAMAI_PRAGMA),
    }
    start = time()
    response = get(url, headers=headers)
    end = time()
    timestamp = datetime.now().isoformat()
    time_taken = '%0.3fs' % (end - start)
    content_length = '%0.2fKB' % (int(response.headers['content-length']) / 1024)
    cache_info = response.headers['X-Cache'].split()
    cache_code = cache_info[0]
    edge_host = cache_info[2]
    headers = pformat(response.headers)
    return timestamp, url, time_taken, content_length, cache_code, edge_host, headers

def collect_stats(edge_hostnames, urls, rounds=1):
    """Collects and saves the stats of HTTP requests made to the provided hosts/urls."""
    # build a list of all urls needed
    urls = ['http://%s%s' % (edge_hostname, url) 
            for edge_hostname in edge_hostnames for url in urls]
    # collect the stats by making http requests 
    stats = [request(url) for url in urls * rounds]

    # write the stats in a csv file 
    now = datetime.now().isoformat()[:16].replace(':', '-')
    csv_file_name = 'clarotests_%s.csv' % now
    with open(csv_file_name, 'wb') as csv_file:
        w = writer(csv_file)
        w.writerows(stats)
    return csv_file_name


if __name__ == '__main__':
    csv_file = collect_stats(EDGE_HOSTNAMES, URLS)
    print 'Stats collected successfully in %s' % csv_file