[Biopython-dev] BioSQL: patch for discussion

[Brad Chapman]

Hi Yves;

> attached is a patch for bringing BioSQL to the current database schema, 
> since Brad seems too busy graduating :)

Heh. Graduating. That would be a nice thing. I have heard that
sometimes they let folks do that :-).

Thanks much for these patches. I'm sorry I haven't had a chance to
look at them earlier -- I have been fighting with this cluster for
so many days and hours it's overall an ugly glazed over look that
fills my eyes. 

I checked in your changes but couldn't fix the Test suite and new
SQL because for some unknown reason I suddenly can't make the test
suite load any SQL without complaining on the comments. I'm sure it
is something foolish, but after a million straight hours on the
computer I can't get it figured. Oh well, that'll get updated soon
enough.

> One of the things to discuss is that I broke the current interface: I 
> added the sequence version to the db items ids (breaking test_BioSQL's 
> assert item_ids == ['AF297471', 'AJ237582', 'L31939', 'M81224', 
> 'X55053', 'X62281']).  I think this is necessary, since accession 
> numbers are not guaranteed unique...
> 
> Another breakage vis-à-vis the test suite is that "no strand info" is 
> dignified by None instead of 0.

Both of these changes make good sense. The test suite should really
be updated to deal with them. I will do that real soon like when I
regain my sanity (or if someone beats me to it :-).

Checking-in-patches-and-making-no-sense-and-heading-to-sleep-ly yr's,
Brad