Grepping around the code I think this is the only way a ftype attribute on an output affects the evaluation of test data. if attributes.get( 'ftype', None ) == 'bam': local_fh, temp_name = self._bam_to_sam( local_name, temp_name ) local_name = local_fh.name I am not sure it was ever meant as a strict test. I worry about breaking backward compatibility but it is easy enough to implement this as an actual check when using newer API driven tests. I have opened a pull request for this functionality here: https://bitbucket.org/galaxy/galaxy-central/pull-request/347/check-ftype-att... Thoughts? -John On Tue, Mar 11, 2014 at 10:30 AM, Peter Cock <p.j.a.cock@googlemail.com> wrote:
On Tue, Mar 11, 2014 at 2:58 PM, Björn Grüning <bjoern.gruening@gmail.com> wrote:
Hi Peter,
I think you need to have:
<change_format> <when input="output.out_format" value="0" format="txt"/> <when input="output.out_format" value="0 -html" format="html"/> <when input="output.out_format" value="2" format="txt"/> <when input="output.out_format" value="2 -html" format="html"/> <when input="output.out_format" value="4" format="txt"/> <when input="output.out_format" value="4 -html" format="html"/> <when input="output.out_format" value="5" format="blastxml"/> </change_format>
Sorry, can't test it right now.
That seems to work - thanks Björn :)
This seems to have exposed a bug in the <test> framework, e.g.
<test> <param name="query" value="rhodopsin_nucs.fasta" ftype="fasta" /> <param name="db_opts_selector" value="file" /> <param name="subject" value="three_human_mRNA.fasta" ftype="fasta" /> <param name="database" value="" /> <param name="evalue_cutoff" value="1e-40" /> <param name="out_format" value="5" /> <param name="adv_opts_selector" value="basic" /> <output name="output1" file="blastn_rhodopsin_vs_three_human.xml" ftype="blastxml" /> </test>
This was saying the output should have been "blastxml", but until I just fixed it the output was being tagged as "tabular" (although run_functional_tests.sh did check the content it didn't check the datatype matched).
Dave - do think this is a reasonable enhancement?
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/