Great solution...
Glad I could help!

On Dec 05, 2012, at 02:56 PM, Nate Coraor <n...@bx.psu.edu> wrote:

On Dec 5, 2012, at 5:10 PM, Anthonius deBoer wrote:

> I confirmed that you have to include the version numbers for these toolshed tools:
>
> This works:
> toolshed.g2.bx.psu.edu/repos/devteam/bwa_wrappers/bwa_wrapper/1.2.3 = drmaa://-V -q long.q -pe smp 4/
>
> this does not:
> toolshed.g2.bx.psu.edu/repos/devteam/bwa_wrappers/bwa_wrapper = drmaa://-V -q long.q -pe smp 4/
>
> Thon

Hi Thon,

I just committed a fix for this in 75a97c971d36 that will be available in our next distribution release. It will also match from most to least specific so you can define all of:

toolshed.g2.bx.psu.edu/repos/devteam/bwa_wrappers/bwa_wrapper/1.2.3 = foo://
toolshed.g2.bx.psu.edu/repos/devteam/bwa_wrappers/bwa_wrapper = bar://
bwa_wrapper = baz://

In this example:

- version 1.2.3 of the bwa wrapper installed from the toolshed would use foo://
- all other versions installed from the toolshed would use bar://
- any tool with <tool id="bwa_wrapper">, regardless of how it's installed (from the toolshed or directly added to tool_conf.xml) would use baz://

Only the most specific match will be used if there are multiple matches (as in the example).

--nate

>
> On Dec 05, 2012, at 02:00 PM, Anthonius deBoer <thondeb...@me.com> wrote:
>
>> Hi Ross,
>>
>> I was indeed expecting something like that and when I tried what you suggestyed, it worked :)
>> you need to use the complete path to the tool (including version number??!?!) in the tool-runner definition to make it be able to be set in the universe_wsgi.ini...
>>
>> So indeed the toolshed move of BWA breaks the individual tool_runner settings and makes for one ugly config file now...
>> But at least we found a workaround...
>>
>> Thanks
>>
>> Thon
>>
>> On Dec 05, 2012, at 01:52 PM, Ross <ross.laza...@gmail.com> wrote:
>>
>>> Hi, Thon
>>> I'm guessing here: but now that bwa is kept in the toolshed, the tool name appears to be toolshed.g2.bx.psu.edu/repos/devteam/bwa_wrappers/bwa_wrapper/1.2.3 according to your job runner entry - so I wonder if that whole ugly string now needs to be in the tool_runners section to over-ride the default rather thant "bwa_wrapper" ?
>>>
>>> This might be an unforseen consequence of the way the toolshed adapts tool names to keep versions under control. Fixing this might be a little tricky since one goal is to keep multiple tool versions separate - but we will all want all versions of bwa_wrapper to use the same runner flags?
>>>
>>>
>>> On Thu, Dec 6, 2012 at 8:40 AM, Anthonius deBoer <thondeb...@me.com> wrote:
>>> Hi,
>>>
>>> In my fresh copy of galaxy-dist I am having a problem with tools being run only with the default tool runner and not with the assigned tool runner I created in the [galaxy:tool_runners] section...
>>>
>>> This is the setting for bwa_wrapper in my universe_wsgi.ini:
>>>
>>> bwa_wrapper = drmaa://-V -q long.q -pe smp 4/
>>>
>>> and this is what the job admin page shows:
>>>
>>> It's in a different queue and does not use the smp setting I ask for...(It's the default setting)
>>>
>>>
>>>
>>> any ideas?
>>>
>>>
>>> 7 X 26 minutes ago toolshed.g2.bx.psu.edu/repos/devteam/bwa_wrappers/bwa_wrapper/1.2.3 running python /mnt/ngs/analysis/svcgalaxy/shed_tools/toolshed.g2.bx.psu.edu/repos/devteam/bwa_wrappers/ffa8aaa14f7c/bwa_wrappers/bwa_wrapper.py --threads="4" --fileSource=indexed --ref="/mnt/ngs/analysis/svcgalaxy/DATA/galaxy-indexes/bwa-indexes/human_g1k_v37_decoy.fasta" --do_not_build_index --input1=/home/tdeboer/P/MUT190-MAIN/fastq-2_205/2_205_1_AAACAT_R1.fastq --input2=/home/tdeboer/P/MUT190-MAIN/fastq-2_205/2_205_1_AAACAT_R2.fastq --output=/mnt/ngs/analysis/svcgalaxy/DATA/files/000/dataset_776.dat --genAlignType=paired --params=pre_set --suppressHeader=false drmaa://-V -q all.q -pe smp 1/ 4440144
>>>
>> ___________________________________________________________
>> 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/
> ___________________________________________________________
> 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/

___________________________________________________________
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/

Reply via email to