Thanks, Alex

I don't think that this works for me because the xml file doesn't know how many output files there will be, nor what they should be named. Only the program executable knows that. If I knew at xml processing time, I could use one of the other flavors of handling multiple output files.

I can encode this information into the <name> field of the prescribed related file name format, and would like that name to be included in the visible history string, or simply be the history string. That way a user will know the contents of each output file / dataset. I think that this change is fairly simple for someone who knows the mechanisms of collect_primary_datasets() and the HistoryDatasetAssociation class. But being my third day looking at Galaxy, I haven't figured it out yet.

-Al

-----Original Message----- From: Bossers, Alex [mailto:Alex.Bossers@wur.nl] Sent: Friday, October 22, 2010 1:49 AM To: Al Simons Cc: galaxy-dev@lists.bx.psu.edu Subject: RE: [galaxy-dev] More on variable multiple output files: identification

Al,

I had a question on this 14th of October as well but didn't get any answer. Actually it was one level further than what you are asking so I think for you the solution can be something like this that allows the change of the history description:

<data name="out_tool" format="text" label="tool ${cmd.value_label}" /> </outputs>

In this example the 'cmd' is a type=select where the option value gets displayed in the history. There are more options to also have the dataset name (using ${input1.name} )and such....but the wiki is not the place to find these answers....

As asked in my mail of Oct 14th I still would like to know how to use this technique if the $cmd example from above is inside a conditional named tool. It would require one extra level in the "array";

<data name="out_tool" format="text" label="tool ${tool.cmd.value_label}" /> </outputs>

In the cheetah script you can easily use #if $tool.cmd=="xx": do something #end if But for the output it doesn't work. Should be something obvious.

Alex

-----Oorspronkelijk bericht----- Van: galaxy-dev-bounces@lists.bx.psu.edu [mailto:galaxy-dev-bounces@lists.bx.psu.edu] Namens Al Simons Verzonden: donderdag 21 oktober 2010 21:40 Aan: galaxy-dev@lists.bx.psu.edu Onderwerp: Re: [galaxy-dev] More on variable multiple output files: identification

By reading the Galaxy code, I've found exec_after_process() and seen examples in various tools, so that answers the "how do you do post processing?" question. However, renaming a history item is still beyond me. Any pointers would be appreciated in how to manipulate the HDA.

Thanks, -Al ________________________________________ From: galaxy-dev-bounces@lists.bx.psu.edu [galaxy-dev-bounces@lists.bx.psu.edu] On Behalf Of Al Simons [Al.Simons@jax.org] Sent: Thursday, October 21, 2010 1:44 PM To: galaxy-dev@lists.bx.psu.edu Subject: [galaxy-dev] More on variable multiple output files: identification

Thanks to the list's help, I have my multiple output files. The next problem is that the displayed descriptions all have the identical text as the "master" (first output) file. I need to know which is which. The data in each file that my tool is indistinguishable, so I can't label the data internally. I need to precisely track which resultant dataset corresponds to which of the program's output streams. I can't even dead-reckon, because the datasets appear to be randomly ordered with respect to either the file's creation time or its original name.

It would be _really_ nice if the output name extracted from the file name could be included in (or just be) the dataset name displayed in the history. That would give the program control of the displayed name, which would be extremely helpful.

I see a tantalizing clue at the bottom of the ToolsMultipleOutput page, to reference "__collected_datasets__['primary'][name]." in a post processing hook. However, I can't find any reference to post processing hooks in the wiki or on Google. I'd welcome a pointer to any documentation about post-processing. Is there a documentation section that I'm overlooking?

Thanks for the support, -Al _______________________________________________ galaxy-dev mailing list galaxy-dev@lists.bx.psu.edu http://lists.bx.psu.edu/listinfo/galaxy-dev

_______________________________________________ galaxy-dev mailing list galaxy-dev@lists.bx.psu.edu http://lists.bx.psu.edu/listinfo/galaxy-dev

What service! MEGAthanks... On to learning about getting the latest from GIT. (smile)

-Al

-----Original Message----- From: Daniel Blankenberg [mailto:dan@bx.psu.edu] Sent: Friday, October 22, 2010 10:19 AM To: Al Simons Cc: Bossers, Alex; galaxy-dev@lists.bx.psu.edu Subject: Re: [galaxy-dev] More on variable multiple output files: identification

Hi Al,

Changeset 4465:78d2a72ee1d6 adds the provided name/designation in parenthesis to the name of the 'base' output dataset, similar to: 'toolName on data x, y (newDatasetName)'. Let us know if this doesn't address your issue.

Thanks for using Galaxy,

Dan

Thanks, Alex

I don't think that this works for me because the xml file doesn't know how many output files there will be, nor what they should be named. Only the program executable knows that. If I knew at xml processing time, I could use one of the other flavors of handling multiple output files.

I can encode this information into the <name> field of the prescribed related file name format, and would like that name to be included in the visible history string, or simply be the history string. That way a user will know the contents of each output file / dataset. I think that this change is fairly simple for someone who knows the mechanisms of collect_primary_datasets() and the HistoryDatasetAssociation class. But being my third day looking at Galaxy, I haven't figured it out yet.

-Al

-----Original Message----- From: Bossers, Alex [mailto:Alex.Bossers@wur.nl] Sent: Friday, October 22, 2010 1:49 AM To: Al Simons Cc: galaxy-dev@lists.bx.psu.edu Subject: RE: [galaxy-dev] More on variable multiple output files: identification

Al,

I had a question on this 14th of October as well but didn't get any answer. Actually it was one level further than what you are asking so I think for you the solution can be something like this that allows the change of the history description:

<data name="out_tool" format="text" label="tool ${cmd.value_label}" /> </outputs>

In this example the 'cmd' is a type=select where the option value gets displayed in the history. There are more options to also have the dataset name (using ${input1.name} )and such....but the wiki is not the place to find these answers....

As asked in my mail of Oct 14th I still would like to know how to use this technique if the $cmd example from above is inside a conditional named tool. It would require one extra level in the "array";

<data name="out_tool" format="text" label="tool ${tool.cmd.value_label}" /> </outputs>

In the cheetah script you can easily use #if $tool.cmd=="xx": do something #end if But for the output it doesn't work. Should be something obvious.

Alex

-----Oorspronkelijk bericht----- Van: galaxy-dev-bounces@lists.bx.psu.edu [mailto:galaxy-dev-bounces@lists.bx.psu.edu] Namens Al Simons Verzonden: donderdag 21 oktober 2010 21:40 Aan: galaxy-dev@lists.bx.psu.edu Onderwerp: Re: [galaxy-dev] More on variable multiple output files: identification

By reading the Galaxy code, I've found exec_after_process() and seen examples in various tools, so that answers the "how do you do post processing?" question. However, renaming a history item is still beyond me. Any pointers would be appreciated in how to manipulate the HDA.

Thanks, -Al ________________________________________ From: galaxy-dev-bounces@lists.bx.psu.edu [galaxy-dev-bounces@lists.bx.psu.edu] On Behalf Of Al Simons [Al.Simons@jax.org] Sent: Thursday, October 21, 2010 1:44 PM To: galaxy-dev@lists.bx.psu.edu Subject: [galaxy-dev] More on variable multiple output files: identification

Thanks to the list's help, I have my multiple output files. The next problem is that the displayed descriptions all have the identical text as the "master" (first output) file. I need to know which is which. The data in each file that my tool is indistinguishable, so I can't label the data internally. I need to precisely track which resultant dataset corresponds to which of the program's output streams. I can't even dead-reckon, because the datasets appear to be randomly ordered with respect to either the file's creation time or its original name.

It would be _really_ nice if the output name extracted from the file name could be included in (or just be) the dataset name displayed in the history. That would give the program control of the displayed name, which would be extremely helpful.

I see a tantalizing clue at the bottom of the ToolsMultipleOutput page, to reference "__collected_datasets__['primary'][name]." in a post processing hook. However, I can't find any reference to post processing hooks in the wiki or on Google. I'd welcome a pointer to any documentation about post-processing. Is there a documentation section that I'm overlooking?

Thanks for the support, -Al _______________________________________________ galaxy-dev mailing list galaxy-dev@lists.bx.psu.edu http://lists.bx.psu.edu/listinfo/galaxy-dev

---

galaxy-dev mailing list galaxy-dev@lists.bx.psu.edu http://lists.bx.psu.edu/listinfo/galaxy-dev

---

galaxy-dev mailing list galaxy-dev@lists.bx.psu.edu http://lists.bx.psu.edu/listinfo/galaxy-dev