[Python-Dev] Summary of Python tracker Issues

[Ezio Melotti]

Hi,

On Fri, Nov 30, 2012 at 11:52 PM, Brett Cannon <brett at python.org> wrote:

> On Fri, Nov 30, 2012 at 4:07 PM, R. David Murray <rdmurray at bitdance.com>wrote:
>
>> On Fri, 30 Nov 2012 14:38:12 -0500, Brett Cannon <brett at python.org>
>> wrote:
>> > Do we have a graph of the historical trend of the number of bugs (or at
>> > least the historical details stored somewhere)? I think we have had a
>> net
>>
>> Not really.  Ezio made one by hand once, but there is nothing automated.
>>
>
The one I made can be found here:
https://docs.google.com/spreadsheet/ccc?key=0AplyAWXqkvHUdFF0SkVrT3VKcnRBZXNrR1hleHowWnc
I now updated it with the latest data.
On the Sheet 2 you can find additional graphs that show the releases of
Python together with the data.  Only final releases are included, alphas,
betas and rcs are not included.
The spreadsheet is a bit messy because I was experimenting with different
kind of graphs and trying to work around some limitations of Google Docs,
but it should be good enough.


>
>> The historical details are stored only in the mailing list archives, as
>> far as I know.  In theory I think you could re-calculate them from the
>> Roundup DB, but for various reasons the numbers would probably come out
>> slightly different.  Still, getting the data from the DB would be better
>> than parsing the emails, since for one reason and another there are
>> missing Friday reports, and reports that were issued on non-Friday
>> dates.
>>
>
One option I was considering is having the weekly report script append the
result on a file and make it available on bugs.python.org, or even use it
to generate graphs directly.  This is something I considered and planned to
implement for a long time, but haven't done it yet.


>> > decrease in open bugs the last couple of weeks and it would be neat to
>> see
>> > an absolute and relative graph of the overall trend since Python 3.3.0
>> was
>> > released. Also might make a nice motivator to try to close issues
>> faster. =)
>> >
>> > Otherwise is the code public for this somewhere? I assume it's making an
>>
>> Yes.  It is in the software repository for our roundup instances:
>>
>>
>> http://hg.python.org/tracker/python-dev/file/default/scripts/roundup-summary
>>
>> (Be warned that that isn't the location from which the script is
>> executed, so it is possible for what is actually running to get out of
>> sync with what is checked in at that location.)
>>
>> > XML-RPC call or something every week to get the results, but if I
>> decide to
>>
>> Nope, it talks directly to the DB.  And as you will see, it is more
>> than a bit gnarly.
>>
>>
> I think I could also download the csv file and parse that to get whatever
> data I wanted.
>

To figure out when an issue was closed you need to access its history, and
that's not available through XML-RPC/csv IIRC.  You should be able to
figure out when the issue got created though.
Anyway, it's probably easier to implement something like what I mentioned
earlier.


>
>>  > do a little App Engine app to store historical data and do a graph I
>> would
>> > rather not have to figure all of this out from scratch. =) Although I
>> could
>> > I guess also parse the email if I wanted to ignore all other emails.
>>
>> I'm not sure how one would go about integrating the above with an App
>> Engine app.  I suspect that not quite enough information is available
>> through the XML-RPC interface to replicate that script, but maybe you
>> could manage just the open-close counting part of it.  I haven't
>> looked at what it would take.
>>
>
> It really depends on what statistics I cared about (e.g. there are less
> than 4000 bugs while there are less than 25,000 closed bugs). If I just did
> high-level statistics it wouldn't be bad, but if I try to track every issue
> independently that might be annoying (and actually cost money for me,
> although I already personally pay for py3ksupport.appspot.com so I can
> probably piggyback off of that app's quota). We will see if this ever goes
> anywhere. =)
>
>
Another somehow related project/experiment I've been working on is
collecting stats about the patches available on the tracker.  I put
together a temporary page that allows you to enter the name of a module (or
any file/path) and get a list of issues with patches that affect the
specified module(s):  http://wolfprojects.altervista.org/issues.html
FTR this is based on the word done by anatoly (see links on the page).
I'm planning to eventually integrate this in the tracker too, but lately I
don't have too much time, so there's no ETA.

Best Regards,
Ezio Melotti
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-dev/attachments/20121201/ea177793/attachment-0001.html>