Hi Peter, What is the test error given when you do have a value defined for name in output? Can you try using 'empty_file.dat'? e.g. <output name="out_file" file="empty_file.dat" > or <output name="out_file" file="empty_file.dat" compare="contains"> etc Thanks for using Galaxy, Dan On Apr 4, 2013, at 1:34 PM, Peter Cock wrote:
On Wed, Apr 3, 2013 at 2:48 PM, Peter Cock <p.j.a.cock@googlemail.com> wrote:
On Wed, Apr 3, 2013 at 2:37 PM, Dave Bouvier <dave@bx.psu.edu> wrote:
For the outputs, rgClean.xml provides an example of comparing a composite output dataset with the expected test data.
Here there is a tag <extra_files> used within the <output> tag.
Many thanks Dave - those should be enough to get me started, and maybe I can update the wiki once I'm happy with how this works...
I'm trying this for makeblastdb which produces a single output, a composite datatype which is a BLAST database.
The examples I found all seemed to be composite datatypes with a central file (e.g. HTML plus images), i.e. composite_type='auto_primary_file'
In the case of the BLAST databases, this is a composite datatype without a primary file, aka composite_type='basic'. Leaving the <param> without a file gives:
Exception: Test output does not have a 'file'
Using a empty value doesn't work,
Traceback (most recent call last): File "/mnt/galaxy/galaxy-central/test/functional/test_toolbox.py", line 171, in test_tool self.do_it( td, shed_tool_id=shed_tool_id ) File "/mnt/galaxy/galaxy-central/test/functional/test_toolbox.py", line 102, in do_it self.verify_dataset_correctness( outfile, hid=elem_hid, maxseconds=testdef.maxseconds, attributes=attributes, shed_tool_id=shed_tool_id ) File "/mnt/galaxy/galaxy-central/test/base/twilltestcase.py", line 828, in verify_dataset_correctness self.files_diff( local_name, temp_name, attributes=attributes ) File "/mnt/galaxy/galaxy-central/test/base/twilltestcase.py", line 66, in files_diff local_file = open( file1, 'U' ).readlines() IOError: [Errno 21] Is a directory: '/mnt/galaxy/galaxy-central/test-data'
Using /dev/null gets a bit further but something is confusing the comparisons.
My current experimental test looks like this:
<tests> <test> <param name="dbtype" value="prot"/> <param name="file" value="four_human_proteins.fasta"/> <param name="title" value="Just 4 human proteins"/> <param name="parse_seqids" value=""/> <param name="hash_index" value="-hash_index"/> <output name="out_file" file="/dev/null" ftype="blastdbp"> <extra_files type="file" value="four_human_proteins.fasta.phd" name="blastdb.pdb"/> <extra_files type="file" value="four_human_proteins.fasta.phi" name="blastdb.phi"/> <extra_files type="file" value="four_human_proteins.fasta.phr" name="blastdb.phr"/> <extra_files type="file" value="four_human_proteins.fasta.pin" name="blastdb.pin"/> <extra_files type="file" value="four_human_proteins.fasta.pog" name="blastdb.pog"/> <extra_files type="file" value="four_human_proteins.fasta.psd" name="blastdb.psd"/> <extra_files type="file" value="four_human_proteins.fasta.psi" name="blastdb.psi"/> <extra_files type="file" value="four_human_proteins.fasta.psq" name="blastdb.psq"/> </output> </test> </tests>
Any advice?
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/