VoIP Mailing List Archives

[lfurrea at gmail.com]

Hi all,

We are writing a xml_cdr parser to load CDRs in SQLite. We are interested in logging times for both A leg and B leg so that transfers are reported as individual calls with accurate timing. eg Inboud call to AA lasted 14 seconds then call to operator 20s and then call to actual extension 5min

As of now we are using the <callflow> tag with the "number" attribute to find out who did the A leg talk to, then we open the B leg files and get the times from each jump from the <times> tag within the <callflow> tag on the B leg file.

Is this right or maybe someone could suggest a better way to do it.

TIA

Luis

[lfurrea at gmail.com]

On our test calls we haven't been able to correlate times from the A leg with times from the B leg.

I would expect something as A-leg(duration)= B-leg1(duration)+B-leg2(duration)

Also the <times> tag within <callflow> tag does not seem to be in epoch microseconds. so it does not seem that's where i should be looking for that info.

Here's an example of the <times> tag for a test call on the A-Leg:

1. <times>
   
2. <created_time>1233942283835696</created_time>
   
3. <profile_created_time>1233942283835696</profile_created_time>
   
4. <progress_time>1233942283999716</progress_time>
   
5. <progress_media_time>1233942283999716</progress_media_time>
   
6. <answered_time>1233942287291931</answered_time>
   
7. <hangup_time>0</hangup_time>
   
8. <transfer_time>1233942303240916</transfer_time>
   
9. </times>
   

any hint is appreciated


On Thu, Feb 12, 2009 at 3:46 PM, Luis F Urrea <lfurrea@gmail.com (lfurrea@gmail.com)> wrote:

[msc at freeswitch.org]

Pastebin the whole file so that we can see it in context...
-MC

On Thu, Feb 12, 2009 at 2:50 PM, Luis F Urrea <lfurrea@gmail.com> wrote:

[lfurrea at gmail.com]

Heres pastebin of the A-leg

http://pastebin.com/m6731913d


On Thu, Feb 12, 2009 at 5:00 PM, Michael Collins <msc@freeswitch.org (msc@freeswitch.org)> wrote:

[msc at freeswitch.org]

On Thu, Feb 12, 2009 at 3:31 PM, Luis F Urrea <lfurrea@gmail.com> wrote:

[lfurrea at gmail.com]

My mistake, they do seem to be microsecs.


But still I cannot correlate times from the A-leg with the B-legs.

I have included below the xml_cdr files generated for the test call.

The test call was made using three registered extensions. Basically, Ext 201 calls ext 203 and they talk, then 203 blindly transfers to 202, 202 does not answer and call rolls to voicemail.

A-leg:
http://pastebin.freeswitch.org/7206

B-leg:
http://pastebin.freeswitch.org/7204

B-leg:
http://pastebin.freeswitch.org/7205


Thanks for your help

On Thu, Feb 12, 2009 at 6:08 PM, Michael Collins <msc@freeswitch.org (msc@freeswitch.org)> wrote:

[anthony.minessale at g...]

each b leg call on the a leg shows up in a <callflow> tag

At the bottom of
http://pastebin.freeswitch.org/7206

<times>
<created_time>1233942283835696</created_time>
<profile_created_time>1233942283835696</profile_created_time>
<progress_time>1233942283999716</progress_time>
<progress_media_time>1233942283999716</progress_media_time>
<answered_time>1233942287291931</answered_time>
<hangup_time>0</hangup_time>
<transfer_time>1233942303240916</transfer_time>
</times>

note transfer time, this is the time that the call was transferred to another extension.

<transfer_time>1233942303240916</transfer_time>

divide this number by one million to get epoch time

1233942303240916 / 1000000 == 1233942303


now look at http://pastebin.freeswitch.org/7204 at the bottom
<times>
<created_time>1233942303325062</created_time>
<profile_created_time>1233942303325062</profile_created_time>
<progress_time>1233942303368916</progress_time>
<progress_media_time>0</progress_media_time>
<answered_time>0</answered_time>
<hangup_time>1233942333010768</hangup_time>
<transfer_time>0</transfer_time>
</times>


This is the b leg, see the created_time:

<created_time>1233942303325062</created_time>

1233942303325062 / 1000000 == 1233942303


so as you can see the epoch time of your b leg cdr has a created_time that is the same one second window that corresponds to the transfer_time in the callflow tag in your a leg cdr


On Fri, Feb 13, 2009 at 8:42 AM, Luis F Urrea <lfurrea@gmail.com (lfurrea@gmail.com)> wrote: