on-net-sched.py

import os
import requests
from time import asctime,localtime,sleep
import sys
if len(sys.argv)==3:
    command=sys.argv[1]
    sleeptime=int(sys.argv[2])*60
elif len(sys.argv)==2:
    command=sys.argv[1]
    sleeptime=30*60 # default 30 mins
else:
    print "USAGE:  ",os.path.basename(sys.argv[0]),' "<command>" <sleeptime(optional)>'
    sys.exit()
connected=False
count=1
while not connected:
    print 'attempt #',count,' at ',asctime(localtime())
    count+=1
    try:
        r=requests.get('http://www.paranatural.net/') # test page to get which is known to work with default user agent
    except:     
        sleep(sleeptime) #sleep for an hour
        continue
    connected=r.status_code == 200
else:
    print 'success!! attempt no. #',count,' at ',asctime(localtime())
    os.system(command)

sample usage

H:\droidtut>on-net-sched.py "youtube-dl https://www.youtube.com/playlist?list=PLB03EA9545DD188C3" 1
attempt # 1  at  Thu Oct 23 15:48:02 2014
attempt # 2  at  Thu Oct 23 15:49:02 2014
attempt # 3  at  Thu Oct 23 15:50:03 2014
success!! attempt no. # 4  at  Thu Oct 23 15:50:04 2014
[youtube:playlist] PLB03EA9545DD188C3: Downloading webpage
[download] Downloading playlist: Android Development For Absolute Beginners
[youtube:playlist] playlist Android Development For Absolute Beginners: Collected 40 video ids (downloading 40 of them)
....

Nice one, but instead of GET I would use HEAD - saves some time and bandwidth .