This only affects tests that are run manually, the automated testing framework correctly resolves tool dependencies.

   --Dave B.

On 5/9/13 10:41:22.000, Peter Cock wrote:
On 5/8/13 20:08:08.000, Ira Cooke wrote:

Hi All,

I'm starting to add functional tests to my tools but I've found that I
can't get the test framework to work well with tool dependencies.

The tool I'm considering is xtandem

http://testtoolshed.g2.bx.psu.edu/view/iracooke/xtandem

When I install this tool on my local galaxy instance (on a clean install)
the tool runs fine and if I look in the handler logs I see

executing:
PACKAGE_BASE=/home/iracooke/tool_dependencies/galaxy_protk/1.2.0/iracooke/xtandem/0940ff5d28ab;
export PACKAGE_BASE; .
/home/iracooke/tool_dependencies/galaxy_protk/1.2.0/iracooke/xtandem/0940ff5d28ab/env.sh;
PACKAGE_BASE=/home/iracooke/tool_dependencies/trans_proteomic_pipeline/4.6.1/iracooke/xtandem/0940ff5d28ab;
export PACKAGE_BASE; .
/home/iracooke/tool_dependencies/trans_proteomic_pipeline/4.6.1/iracooke/xtandem/0940ff5d28ab/env.sh;
rvm 1.9.3@protk-1.2.0 do tandem_search.rb -d
/home/iracooke/galaxy-central/database/files/000/dataset_1.dat
--var-mods='            57.021464__at__C                '       --fix-mods='
15.994915__at__M                '
/home/iracooke/galaxy-central/database/files/000/dataset_2.dat -o
/home/iracooke/galaxy-central/database/files/000/dataset_4.dat -r
--enzyme=Trypsin --precursor-ion-tol-units=ppm -v 0 -f 0.65 -p 100.0
--strict-monoisotopic-mass --keep-params-files
galaxy.jobs DEBUG 2013-04-27 06:53:35,586 (4) Persisting job destination
(destination id: local:///)


Note that in this case the tool environments from my tool dependencies are
loaded ... and the tool runs fine.

When running functional tests however the tests are executed without first
loading the environments from tool dependencies ... and naturally the tool
fails.  On my local machine I get the following when running the tests (Note
that no loading of dependency shel environments occurs).  (This seems to be
the same issue on the testtoolshed.g2.bx.psu.edu ).

Job dispatched
galaxy.jobs.runners.local DEBUG 2013-04-27 06:55:19,564 (3) executing: rvm
1.9.3@protk-1.2.0 do tandem_search.rb -d
/tmp/tmp3I8GgN/database/files/000/dataset_1.dat             --var-mods='
15.994915__at__M                ,               '               --fix-mods='
57.021464__at__C                '
/tmp/tmp3I8GgN/database/files/000/dataset_2.dat -o
/tmp/tmp3I8GgN/database/files/000/dataset_3.dat -r --enzyme=Trypsin
--precursor-ion-tol-units=ppm -v 0 -f 0.65 -p 100.0
--strict-monoisotopic-mass --keep-params-files
galaxy.jobs DEBUG 2013-04-27 06:55:19,730 (3) Persisting job destination
(destination id: local:///)
galaxy.jobs.runners.local DEBUG 2013-04-27 06:55:19,889 execution
finished: rvm 1.9.3@protk-1.2.0 do tandem_search.rb -d
/tmp/tmp3I8GgN/database/files/000/dataset_1.dat                --var-mods='
15.994915__at__M                ,'              --fix-mods='
57.021464__at__C                '
/tmp/tmp3I8GgN/database/files/000/dataset_2.dat -o
/tmp/tmp3I8GgN/database/files/000/dataset_3.dat -r --enzyme=Trypsin
--precursor-ion-tol-units=ppm -v 0 -f 0.65 -p 100.0
--strict-monoisotopic-mass --keep-params-files
...
galaxy.jobs DEBUG 2013-04-27 06:55:24,277 setting dataset state to ERROR


Is there something wrong with the way I've defined my tool dependencies
for this tool?  Or perhaps it's a bug in the testing framework.

Cheers
Ira

On Thu, May 9, 2013 at 3:10 PM, Dave Bouvier <d...@bx.psu.edu> wrote:
Ira,

There is a known issue with resolving tool dependencies when running
functional tests. One workaround is to set the GALAXY_TOOL_DEPENDENCY_DIR
environment variable to the path configured in your universe_wsgi.ini.

This issue is documented in https://trello.com/c/eVT3QWJF

    --Dave B.

Does this issue affect the Tool Shed tests as well, or just when run
via run_functional_test.sh? (I'm working though a couple of test
failures related to the dependency installation breaking somehow).

Thanks,

Peter

___________________________________________________________
Please keep all replies on the list by using "reply all"
in your mail client.  To manage your subscriptions to this
and other Galaxy lists, please use the interface at:
 http://lists.bx.psu.edu/

To search Galaxy mailing lists use the unified search at:
 http://galaxyproject.org/search/mailinglists/

Reply via email to