[Bioperl-l] Bio::*Taxonomy* changes
cjfields at uiuc.edu
Tue Jul 18 16:05:48 UTC 2006
> I've just realised that there are currently no tests for
> Bio::DB::Taxonomy::flatfile, and that the ones for entrez get skipped.
> Node doesn't get an especially thorough work-out either (in the skipped
> I'm guessing it's not feasible to include the full taxdump from NCBI
> (~40MB) in t/data... do people think it would be reasonable to create
> some sort of small subset of the data? I could just pull out the lines
> from names.dmp and nodes.dmp relevant to a few example organisms. Say,
> for human and a tricky bacteria and virus?
> For the purposes of running the test, where should the index files be
> kept? In t/data with the .dmp files or in /tmp? Should the test script
> delete them afterwards, or leave them be?
I would place a small section in t/data or several individual examples in a
subdirectory thereof (t/data/taxonomy).
> The entrez tests are skipped to 'avoid blocking', but the test only
> makes 2 entrez queries with a sleep(3) in-between. Basically, I don't
> think there's ever any reason to skip. Shall I remove the skip? Lots of
> other database-accessing tests in the test suite just go right ahead and
> access their database, no problem.
Depends on whether there is someone out there who doesn't have a network
connection (and there always is). The DB.t tests skip based on testing for
the env. variable BIOPERLDEBUG.
ok 1 # Skipping tests which require remote servers - set env variable
BIOPERLDEBUG to test
You could always do something along those lines or add a test for a network
connection using an eval block and skip the tests if the network test fails,
but there you run the risk of the tests failing not b/c of code problems but
from remote server issues; I've seen this happen with SwissProt and GenBank
testing before during peak hours.
> Bioperl-l mailing list
> Bioperl-l at lists.open-bio.org
More information about the Bioperl-l