Pycurl: how to determine duration of a request

Question

My task is to do periodic requests to server and write time of this request to database. I use Python 2.7.3, cURL and pycurl as wrapper for cURL. I do request like so:

import pycurl
c = pycurl.Curl()
c.setopt(c.URL, "http://google.com/")
c.setopt(c.VERBOSE, True) # to see request details
c.perform()

But how I can determine the time of this request?

UPD1

Ok, I understand, that I should take 2 timestamps and the diff between them will be the duration of request. But I face some problem:

when I execute:

import pycurl, time
c = pycurl.Curl()
c.setopt(c.URL, "http://google.com/")
c.setopt(c.VERBOSE, True) # to see request details
ts = time.clock()
c.perform()
print time.clock() - ts

I get 0.0 o (sometimes) 0.01. It's wrong diff, because of I execute this commands in python shell and some time is left after I do ts = time.clock() and before I do print time.clock() - ts , so the diff should be about 3 sec.

This output I get on my Ubuntu server 12.04 installed on Virtualbox. Virtualbox installed on Windows 7. When I try code below in windows 7 I get correct output.

Here is another question - maybe I should use time.time() instead of time.clock()?

Solution

Idea was taken from this post . Succinctly, you should use time.clock() on windows and time.time() on Linux or Unix (Ubuntu and others, FreeBSD and others, MacOS). And you can also use timeit.default_timer(). This function detect os type and choose time.time() or time.clock(). So the solution's code is:

import pycurl, timeit
c = pycurl.Curl()
c.setopt(c.URL, "http://google.com/")
c.setopt(c.VERBOSE, True) # to see request details
ts = timeit.default_timer()
c.perform()
print timeit.default_timer() - ts

Answer 1

import pycurl, time
c = pycurl.Curl()
c.setopt(c.URL, "http://google.com/")
c.setopt(c.VERBOSE, True) # to see request details
ts = time.clock()
c.perform()
print time.clock() - ts

Answer 2

don't use time.clock() for timing elapsed intervals. It's also now deprecated.

Answer 3

After you do curl.perform() you can extract information from a curl handle:

m = {}
m['total-time'] = curl.getinfo(pycurl.TOTAL_TIME)
m['namelookup-time'] = curl.getinfo(pycurl.NAMELOOKUP_TIME)
m['connect-time'] = curl.getinfo(pycurl.CONNECT_TIME)
m['pretransfer-time'] = curl.getinfo(pycurl.PRETRANSFER_TIME)
m['redirect-time'] = curl.getinfo(pycurl.REDIRECT_TIME)
m['starttransfer-time'] = curl.getinfo(pycurl.STARTTRANSFER_TIME)

Docs for getinfo(): http://pycurl.io/docs/latest/curlobject.html#pycurl.Curl.getinfo

List of available information to get from request: https://curl.haxx.se/libcurl/c/curl_easy_getinfo.html