[Teammetrics-discuss] The Database Phase

Andreas Tille andreas at an3as.eu
Fri Jun 17 11:57:36 UTC 2011


On Fri, Jun 17, 2011 at 03:26:13AM +0530, Sukhbir Singh wrote:
> Please run archives.sql before trying out the script as it will create
> the database.

Sure I did.  After running the script I've got:

2011-06-17 13:43:12,201 INFO:           Starting ListStat
2011-06-17 13:43:12,232 ERROR: Please set a value for DATABASE username

> You can have a look at the schema also and check whether it is
> according to the way you desired.

I think the schema is OK and your decision to store the hostname in
addition to the mailing list name seems OK to me. 

To your question what to do next there are five options:

  * Bug listmaster at lists.d.o again for providing mbox files or
    suggesting a better solution.
  * Start parsing Vcs commits of teams (Debian Med has SVN and Git)
  * Check my code for Uploaders statistics in UDD
  * Create an output file which can be feed into the R script to
    draw a graph.  The format of the output files can be seen at

       http://blends.debian.net/liststats/*.txt

    Please note that I have basically two algorithms to create such
    a file.  One is in

       svn://svn.debian.org/svn/blends/blends/trunk/team_analysis_tools/archives.sql

    which is not really good and flexible.  An alternative was 
    makes use of crosstab (in postgres-contrib) and should be
    also perfectly applicable to the liststat table.  Have a
    look into upload_history.py and those functions which are
    defined in create_names_prefered.sql.  This should give you
    an idea how to avoid this complex string manipulation stuff
    I did in archives.sql.
  * Try to deal with people posting under different names.

Hope this is enough work for tonight. :-)

Kind regards

        Andreas.

-- 
http://fam-tille.de



More information about the Teammetrics-discuss mailing list