No, I call the executable directly from the xml. It kept failing although it seemed to finish the job and I realized that on success the executable prints a summary message to stderror, because it's using stdout for the actual results so they can be pipped to another command.
From the July news brief I understood that you can now add code to the xml to check the stdout and std error for messages and I was wondering if you have any documentation or sample code for that in any published tool.
Thanks
Birgit Crain, Ph.D. | Sr. Professional Services Scientist | Complete Genomics, Inc. (650) 428-6023 office | (408) 605-3938 mobile bcrain@completegenomics.commailto:bcrain@completegenomics.com
From: John Patterson <jmpatt4@g.uky.edumailto:jmpatt4@g.uky.edu> Date: Thursday, August 16, 2012 12:29 PM To: "galaxy-dev@lists.bx.psu.edumailto:galaxy-dev@lists.bx.psu.edu" <galaxy-dev@lists.bx.psu.edumailto:galaxy-dev@lists.bx.psu.edu> Subject: Re: [galaxy-dev] code example for improved error handling
On 08/16/2012 03:02 PM, Birgit Crain wrote: Hi
I'm writing a tools for an executable that writes results to stdout and reports success message to stderr, so it really need the improved error handling. Is there any sample code or documentation available?
Thanks
Birgit Crain, Ph.D. | Sr. Professional Services Scientist | Complete Genomics, Inc. (650) 428-6023 office | (408) 605-3938 mobile bcrain@completegenomics.commailto:bcrain@completegenomics.com
From: Nicole Rockweiler <n.rockweiler@gmail.commailto:n.rockweiler@gmail.com> Date: Friday, July 20, 2012 10:53 AM To: Brian Haas <bhaas@broadinstitute.orgmailto:bhaas@broadinstitute.org> Cc: "galaxy-dev@bx.psu.edumailto:galaxy-dev@bx.psu.edu" <galaxy-dev@bx.psu.edumailto:galaxy-dev@bx.psu.edu> Subject: Re: [galaxy-dev] pipeline execution succeeds but galaxy shows failure
Hi Brian,
A couple of days ago, smcmanushttps://bitbucket.org/smcmanus pushed the following change to the repo:
Tools can now specify their own handling of stderr and stdout regular expressions as well as exit code ranges.
https://bitbucket.org/galaxy/galaxy-central/issue/325/allow-tool-authors-to-...
It looks like the documentation has yet to be written.
Hope this helps, Nicole
On Fri, Jul 20, 2012 at 12:46 PM, Brad Chapman <chapmanb@50mail.commailto:chapmanb@50mail.com> wrote:
Brian;
I wrote a pipeline (xml attached) that, from what I can gather, succeeds, but galaxy shows it as an error and doesn't make the output file accessible as a new data set.
Is it possible the software is writing to standard error? Galaxy doesn't check status codes, but rather check for stderr and assumes that output indicates a problem. You can wrap the problematic programs with a little script to eat up stderr and check that everything is okay:
http://wiki.g2.bx.psu.edu/Future/Job%20Failure%20When%20stderr
Brad
From the server log, I can see that the command line is being
constructed correctly, and it even indicates that it's captured the output, but in the display of the web browser, it just shows up in the error state. The script being run exits (0) on success. Any ideas?
Here's what the output section of my xml file looks like:
<outputs> <data format="bam" name="coordSortedBam" label="${tool.name<http://tool.name>} on ${on_string}: coord-sorted read alignments" from_work_dir="alignment/alignment.coordSorted.bam"/> </outputs>
and here's what the server log states:
galaxy.jobs.handler INFO 2012-07-20 09:52:05,240 (30) Job dispatched galaxy.jobs.runners.local DEBUG 2012-07-20 09:52:05,453 executing: alignReads.pl --target /Users/bhaas/BioIfx/Galaxy/galaxy-dist/database/files/000/dataset_26.dat -o alignment --aligner bowtie --single /Users/bhaas/BioIfx/Galaxy/galaxy-dist/database/files/000/dataset_23.dat --seqType fq
galaxy.jobshttp://galaxy.jobs DEBUG 2012-07-20 09:52:16,673 finish(): Moved /Users/bhaas/BioIfx/Galaxy/galaxy-dist/database/job_working_directory/000/30/alignment/alignment.coordSorted.bam to /Users/bhaas/BioIfx/Galaxy/galaxy-dist/database/files/000/dataset_50.dat as directed by from_work_dir
Again, as far as I can tell, everything worked - but the browser doesn't think so.
I've run the exact command above on the command-line, and it exits(0) indicating success. Also, I've verified that when run through my galaxy instance, the galaxy-relocated output file is as expected.
Many thanks for your help. I'm still getting my feet wet with galaxy, reading through all the documentation and searching the mailing list for additional help.
best regards,
-brian
--
Brian J. Haas Manager, Genome Annotation and Analysis, Research and Development The Broad Institute http://broad.mit.edu/~bhaashttp://broad.mit.edu/%7Ebhaas
<tool id="alignreads" name="alignReads" version="0.0.1">
<description>alignReads: short read alignment tool wrapper</description> <requirements> <requirement type="package">trinity</requirement> </requirements> <command> alignReads.pl --target $target -o alignment --aligner $aligner_selection.aligner ## Inputs. #if str($inputs.paired_or_single) == "paired": --left $inputs.left_input --right $inputs.right_input #if $inputs.left_input.ext == 'fa': --seqType fa #else: --seqType fq #end if #if str($inputs.library_type) != "None": --SS_lib_type $inputs.library_type #end if --max_dist_between_pairs $inputs.max_dist_between_pairs #else: --single $inputs.input #if str($inputs.input.ext) == 'fa': --seqType fa #else: --seqType fq #end if #if str($inputs.library_type) != "None": --SS_lib_type $inputs.library_type #end if #end if ## Additional parameters. ##if str($inputs.use_additional) == "yes": ## -- $inputs.additional_params ##end if ## direct to output ## > $trinity_log 2>&1 </command> <inputs> <param format="fasta" name="target" type="data" label="target" help="Fasta sequences targeted for short-read alignment" /> <conditional name="inputs"> <param name="paired_or_single" type="select" label="Paired or Single-end data?"> <option value="paired">Paired</option> <option value="single">Single</option> </param> <when value="paired"> <param format="fasta,fastq" name="left_input" type="data" label="Left/Forward strand reads" help=""/> <param format="fasta,fastq" name="right_input" type="data" label="Right/Reverse strand reads" help=""/> <param name="library_type" type="select" label="Strand-specific Library Type"> <option value="None">None</option> <option value="FR">FR</option> <option value="RF">RF</option> </param> <param name="max_dist_between_pairs" type="integer" value="2000" min="1" label="max_dist_between_pairs" help="Maximum length expected between fragment pairs as aligned to the target, including introns where relevant."/> </when> <when value="single"> <param format="fasta,fastq" name="input" type="data" label="Single-end reads" help=""/> <param name="library_type" type="select" label="Strand-specific Library Type"> <option value="None">None</option> <option value="F">F</option> <option value="R">R</option> </param> </when> </conditional> <conditional name="aligner_selection"> <param name="aligner" type="select" label="Select alignment tool to run"> <option value="bowtie">bowtie</option> <option value="bwa">bwa</option> <option value="blat">blat</option> </param> <when value="blat"> <param name="max_intron_length" type="integer" value="10000" min = "1" label="maximum intron length" help="" /> <param name="min_percent_identity" type="integer" value="95" min="1" label="minimum percent identity" help="" /> </when> <when value="bwa"> </when> <when value="bowtie"> </when> </conditional> <!-- <conditional name="use_additional_params"> <param name="use_additional" type="select" label="Use Additional Params?"> <option value="no">No</option> <option value="yes">Yes</option> </param> <when value="no"> </when> <when value="yes"> <param name="additional_params" type="text" value="" label="Additional command-line parameters to aligner" help="" /> </when> </conditional> --> </inputs> <outputs> <data format="bam" name="coordSortedBam" label="${tool.name<http://tool.name>} on ${on_string}: coord-sorted read alignments" from_work_dir="alignment/alignment.coordSorted.bam"/> <!-- <data format="bam" name="nameSortedBam" label="${tool.name<http://tool.name>} on ${on_string}: name-sorted read alignments" from_work_dir="alignment/alignment.nameSorted.bam"/> --> </outputs> <tests> </tests> <help> .. _Trinity: http://trinityrnaseq.sourceforge.net </help></tool> ___________________________________________________________ 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:
___________________________________________________________ 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:
-- Nicole Rockweiler Genome Technology Access Center Washington University in St. Louis Campus Box 8510 4444 Forest Park Avenue Saint Louis, MO 63108
________________________________
The contents of this e-mail and any attachments are confidential and only for use by the intended recipient. Any unauthorized use, distribution or copying of this message is strictly prohibited. If you are not the intended recipient please inform the sender immediately by reply e-mail and delete this message from your system. Thank you for your co-operation.
___________________________________________________________ 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:
Hi
Let me get this straight first. Are you writing python scripts that pass commands to this executable? If so, you are doing something similar to my tool. In this case, all prints to stderr will cause galaxy to report a failed run even if the run is successful.
Set up the command and then make a sub process call with try/except: Need to import errno.
try: retcode = subprocess.call(dividereads_cmd, stderr=subprocess.STDOUT, stdout=splice_log) except OSError, o: if o.errno == errno.ENOTDIR or o.errno == errno.ENOENT: print >> sys.stdout, fail_str, "Error: /insert name here/ not found on this system" exit(1)
Here is some documentation on errno: http://docs.python.org/library/errno.html
The important thing is to redirect the stderr from the executable with the subprocess call. I'd suggest printing your success messages and or debugging to a log file.
-John
________________________________
The contents of this e-mail and any attachments are confidential and only for use by the intended recipient. Any unauthorized use, distribution or copying of this message is strictly prohibited. If you are not the intended recipient please inform the sender immediately by reply e-mail and delete this message from your system. Thank you for your co-operation.