[Biopython-dev] [Bug 2840] When a record has been loaded from BioSQL, trying to save it to another database fails with loader db_loader.load_seqrecord in _load_reference

[bugzilla-daemon at portal.open-bio.org]

http://bugzilla.open-bio.org/show_bug.cgi?id=2840





------- Comment #6 from biopython-bugzilla at maubp.freeserve.co.uk  2009-06-01 11:18 EST -------
Hi David,

I was able to reproduce this problem. When working on Bug 2838, as my test case
I was using just the file cor6_6.gb which by chance has simple reference
locations - and that worked. I have now tested with GI 28804743. Also, using
some of the other GenBank files in our test suites also shows the reference
location problem from BioSQL/Loader.py function _load_reference:

ValueError: invalid literal for int() with base 10: 'None'

This is now fixed in CVS, plus there are now additional unit tests.  For the
fix, I have used a slight variation of Cymon's patch. Does this look sensible
Cymon?

BioSQL/BioSeq.py revision: 1.37
Tests/test_BioSQL.py revision: 1.39
Tests/seq_tests_common.py revision: 1.2

If you could retest with a clean checkout from CVS/github, to confirm the
problem is fixed, that would be great David.

Note - currently in BioSQL we only store one reference location, while GenBank
files can have a single reference covering multiple regions of the record. This
is a limitation of the current BioSQL schema (although it would be interesting
to see how BioPerl deals with this).

Note - there are four known failures in test_BioSQL.py right now, a mixed
strand feature in NC_000932.gb (which triggers two failures), the project cross
reference in NC_005816.gb, and a sub-feature location reference in one_of.gb --
these are all unrelated to this issue (Bug 2840).

Thanks,

Peter


-- 
Configure bugmail: http://bugzilla.open-bio.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.