[Teammetrics-discuss] Fixed bugs in NNTPstat.py

[Andreas Tille]

On Thu, Aug 04, 2011 at 02:07:34AM +0530, Sukhbir Singh wrote:
>     git pull
> 
> There were some bugs in NNTPstat.py that were causing problems. This
> was because I forgot that the liststat code exists after even one team
> is parsed.

This is what I expected to be the reason.
 
> Anyways, I have changed some other things also, most notably you will
> see a more detailed log output as the messages are being downloaded.
> This is the only thing I could come up with, so feel free to comment.

The log is fine.

> I recommend you run the code with two small teams, just to be sure:
> `debian-accessibility` and `debian-blends` (a total of ~6000
> articles). It seems to work for me now but as usual, I leave the
> testing to you :) I think all is well now.

teammetrics=# SELECT project, count(*) from listarchives where domain='lists.debian.org' group by  project;
       project        | count 
----------------------+-------
 debian-amd64         |     5
 debian-curiosa       |  3908
 debian-boot          |   393
 debian-ctte          |     4
 debian-arm           |  9666
 debian-derivatives   |   546
 debian-blends        |  2930
 debian-accessibility |  3177
 debian-desktop       |  2690

Looks good except for amd64 and ctte.  While I think the mboxes have
a reasonable size to reflect the expected number of postings:

/var/cache/teammetrics$ ls -l debian-amd64-2-27592.mbox debian-ctte-1-2608.mbox 
-rw-rw-r-- 1 tille tille 50053394  3. Aug 13:10 debian-amd64-2-27592.mbox
-rw-rw-r-- 1 tille tille  6224948  3. Aug 19:44 debian-ctte-1-2608.mbox

something is wrong with the import.  Could you please have a look into
these two specific lists. 

> Just in case if we run into any more privacy issues (!), we can have
> this code as backup and that is why it should work perfectly.

That's correct.

Kind regards

       Andreas.

-- 
http://fam-tille.de