Re: [GsoC] your latest commit of tasks_udd.py
Hi Akshita,
On Sun, Jul 19, 2015 at 02:33:14AM +0530, Akshita Jha wrote:
>
> I have committed blendstasktools_udd.py and tasks_udd.py to the repository
> with changes.
Fine. I pulled this commit.
> I was unable to test the query "query_new" as the udd.sql.gz
> dump I had downloaded earlier, did not have any records in new_packages
> table.
There was a time when the ftpnew importer was broken. BTW, you can
fix this also locally by cloning
ssh://git.debian.org/git/collab-qa/udd.git
and run
./update-and-run.sh ftpnew
This imports the current new packages.
> I tested the script in yesterdays' (July 18, 2015) sql dump. The
> script "tasks_udd.py" runs without any error for debian-med,
I checked on my side and believe it or not I'm beaten again by some
encoding problem:
$ tasks_udd.py debian-med
Traceback (most recent call last):
File "./tasks_udd.py", line 183, in <module>
print >> f, template.generate(**data).render('xhtml')
UnicodeEncodeError: 'ascii' codec can't encode character u'\xe4' in position 1275: ordinal not in range(128)
I admit I do not have the latest UDD clone but it should not happen also
in older ones. The problem seems to occure in the German index.de.html
since this is the youngest file in the output dir and has 0 Bytes.
Since the english files are created.
I inspected the english version and noticed the following issues:
1. After the last "regular" package that should show up in the bio task
(zalign) the list is continued with "amos-assembler" and this is
displayed three times! This package and the following ones (apollo,
asmlib, ...) should go into the VCS group (please compare with
http://http://blends.debian.org/med/tasks/bio). So thes files need to
be put in a different section and each package should be displayed only
once.
2. In the section "Official Debian packages with lower relevance" that
should end with the package "r-cran-boolnet" another group of packages
is appended. For reasons I do not understand it starts with a triply
copy of aegean (which is in task bio) but a bit later also gnuhealth
from task his and libbio-das-perl from bio-dev are listed. That's wrong
and these packages should not show up here.
3. Again in the section "Debian packages in contrib or non-free" some
random packages are mentioned for no visible reason three times but
here also packages are missing. It should show abyss, arb and others
but shows mozilla-biofox - which is actually tagged "Ignore" and thus
should not be listed at all.
> debian-science
Here I get:
$ tasks_udd.py debian-science
Traceback (most recent call last):
File "./tasks_udd.py", line 183, in <module>
print >> f, template.generate(**data).render('xhtml')
UnicodeEncodeError: 'ascii' codec can't encode character u'\xe4' in position 978: ordinal not in range(128)
Same here - index.de.html has zero bytes.
> and debian-games. The html pages are succesfully created in
> www/<blend>/tasks/ folder.
>
> I noticed that certain information like License for prospective_packages
> was missing from the new html pages that were created. Also, there may be a
> minor issue with the pkgstatus of new packages. I will rectify these issues
> with my next commit.
OK, may be my description above might help to find the problem here.
> If you could go through the code and provide any other feedback, it will be
> very helpful.
For sure I could try to inspect the code for the cause of the problems
above if you want me to do so. The structure of the issues I mentioned
above seem to my experience not really hard to solve but please tell me
explicitely if you need coding help (which I'd gladly provide if you ask
me to).
> Also, what should be our work objective for this week (July 19 - July 25).
Fixing the issues above seems to be a sensible goal. If you manage before
the end of the week (which does not seem improbable to me) we might have
a look into importing the prospective package information from the tasks
files into UDD (which are not yet available there).
Kind regards
Andreas.
--
http://fam-tille.de
Reply to: