September 2012 - galaxy-dev - lists.galaxyproject.org

Mercurial to push changes to toolshed repo
by Joachim Jacob 27 Sep '12

27 Sep '12

Hi all, As a test, I have uploaded some template files to a newly created repository on our test toolshed. I pulled that repository on my development machine using mercurial: from this folder I start now developing the tool wrapper. I have noticed that once the tool name has been set in the tool config (<tool name="xxxxx" >) during the initial upload, pushing a subsequent change to this name, this does not change the name on the repository anymore (and the other settings, for this matter). Is this correct? Thanks, Joachim -- Joachim Jacob, PhD Rijvisschestraat 120, 9052 Zwijnaarde Tel: +32 9 244.66.34 Bioinformatics Training and Services (BITS) http://www.bits.vib.be @bitsatvib

2 3

Best practice: set tool version on wrapper or on tool level?
by Joachim Jacob 27 Sep '12

27 Sep '12

Hi all, I am writing a wrapper around Razers mapper. Currentty, Razers has three versions, the latest being razers3. Since it's under development, what is the best way to deal with new versions? - write a different tool config for every version, giving them different ids like 'razers3', 'razers4'. These tools will appear all simultaneously in the tool box. Every version needs to be installed independently from the toolshed. - or write a tool config with id 'razers' and set a version in the xml <version_command>. Then when a new version comes out, I update the wrapper completely to match the new version. In this way a user can select the different versions can be selected at the top of the tool parameters page in Galaxy, right? How should I deal with naming the version: follow the config file I wrote (version 1.0.0 for razer3, 2.0.0 for razers4 - can cause confusion though.) - or write a wrapper for the tool 'razers' in general, parsing the selection of which version to use to the wrapper, and deal with it over there? I hope I explained my issue clearly. Thanks, Joachim -- Joachim Jacob, PhD Rijvisschestraat 120, 9052 Zwijnaarde Tel: +32 9 244.66.34 Bioinformatics Training and Services (BITS) http://www.bits.vib.be @bitsatvib

2 1

missing .fa files in rsync datacache indexes
by Kevin Y 27 Sep '12

27 Sep '12

Hi, Thanks for providing the indexes via rsync:: http://wiki.g2.bx.psu.edu/Admin/Data%20Integration We've, noticed that despite soft links pointing to it, there are no .fa files for example in hg19/seq/ However there are 2bit files. Would you guys recommend we convert them from 2bit to fasta? Thanks, Kevin.

2 4

[Patch] Interval error
by Björn Grüning 27 Sep '12

27 Sep '12

Hi, for some strange gff files i get the following error: Traceback (most recent call last): File "/media/data/web/galaxy-central/lib/galaxy/jobs/runners/local.py", line 149, in run_job job_wrapper.finish( stdout, stderr, exit_code ) File "/media/data/web/galaxy-central/lib/galaxy/jobs/__init__.py", line 373, in finish dataset.datatype.set_meta( dataset, overwrite = False ) #call datatype.set_meta directly for the initial set_meta call during dataset creation File "/media/data/web/galaxy-central/lib/galaxy/datatypes/interval.py", line 829, in set_meta if valid_start and valid_end and start < end and strand in self.valid_gff3_strand and phase in self.valid_gff3_phase: UnboundLocalError: local variable 'start' referenced before assignment start and end are not defined when elems[3] and elems[4] are '.' Attached is a patch that should fix it. Cheers, Bjoern

1 0

Renaming in Workflows
by Kohler Manuel 26 Sep '12

26 Sep '12

Hi, I have workflows where I would like to rename the item at each step. The reason for this to have a proper naming. What I did is to use the renaming option in the workflow editor. I specified the 'New output name:' to ${input} This does not work. Is there any variable which works or is there a easy way in doing this? Any help or hint is appreciated. Cheers Manuel -- Manuel Kohler Center for Information Sciences and Databases (C-ISD) Department of Biosystems Science & Engineering (D-BSSE) ETH Zurich, Maulbeerstrasse (1078, 1.02), CH-4058 Basel, +41 61 387 3132

2 4

Boolean logic in filters for data in outputs
by Oleksandr Moskalenko 26 Sep '12

26 Sep '12

Hi, I am working on a wrapper that has multiple combinations of switches that lead to the creation of the output files. For instance, one of the files can be generated by conditions "A and B" and also by "A and C". FIrst, I tried the xml <and> and <or> elements, but that didn't work. Next, I tried having two <data> elements with the same name that contained those two pairs of filters. However, it appears that Galaxy only checks the last <data> element, which in this case is "A and C", so "A and B" is never triggered. In other cases there are three variables in play as the two main variables are not initialized if the advanced options conditional parameter's value is on. Passing the options to the wrapper is not a problem - only setting the outputs. I pasted the complete tool definition file at http://pastebin.com/BGDzcc28. Thanks, Alex

1 0

major database corruption
by Julian Pradinuk 26 Sep '12

26 Sep '12

Please help! I'm new to sys admin and I've really messed up. On our server we have a development instance and a production instance of Galaxy. Both run under the same user (galaxy) and each has its own postgresql table (galaxydb_dev and galaxydb_prod). I was testing the ability to backup and restore with the development instance but made the following huge mistake: - ran "pg_dump galaxydb_dev 25_09_2012.sql" - connected to the production instance database by accident - ran "galaxydb_prod=# \i 25_09_2012.sql" Is there any way that I can at least get the important relations and tuples out and in to a new production instance? I have the dump file and the output from running the \i command... Has anyone recovered from a similar situation?

1 0

cufflinks and global_model.txt
by wormbuff 25 Sep '12

25 Sep '12

I recently installed Galaxy on Ubuntu 12.04.1 LTS. Bowtie, SAMtools, and Tophat all run fine with my small subset of paired-end Illumina reads. Unfortunately, it seems to only partially work when I run Cufflinks v2.0.2. Three entires in my history appear (gene expression, transcript expression, and assembled transcripts) containing reasonable output. Each of these entries is highlighted red, though without the red circle with an X. The box contains the message: An error occurred running this job: cufflinks v2.0.2 cufflinks -q --no-update-check -l 300000 -F 0.100000 -j 0.150000 -p 4 cp: cannot stat ` /home/galaxy/galaxy-dist/database/job_working_directory/000/20/global_model.txt': No such file or directory What is this global_model.txt file and why might it be missing/have failed? (sorry about the repost... original got embedded in an unrelated thread)

2 1

where are the tools?
by David Hoover 25 Sep '12

25 Sep '12

I just recently updated to the latest version of galaxy-central, and when I started galaxy, the tools were not listed on the left panel. I could rerun old jobs from my history, however. Why are the tools not listed? David Hoover Helix Systems Staff

2 1

Stop Jobs
by Julian Pradinuk 25 Sep '12

25 Sep '12

Hi all, I've been poking around with the Admin Manage jobs functions - Stop Jobs does not seem to be working. I looked at the code and added some debug prints. When I submit a Stop Jobs in the manager, I don't think the job is making its way to the stop_jobs_queue in the handler.... Are others having issues with this? Is there a configuration I could be missing? Cheers, julian

1 0