Sarah Diehl wrote:

Hi all,

I'm currently working on setting up a fresh Galaxy server at our institute. I'm going through the tool dependencies list and install everything on it. I installed the latest R version (2.12.2) and now I'm trying to install rpy. With some workarounds I managed to compile rpy, but when I try to import it in python, I get the following error:

...

...

...

import rpy

Traceback (most recent call last): File "<stdin>", line 1, in <module> File "rpy.py", line 134, in <module> """ % RVERSION) RuntimeError: No module named _rpy2122

  RPy module can not be imported. Please check if your rpy
  installation supports R 2.12.2. If you have multiple R versions
  installed, you may need to set RHOME before importing rpy. For
  example:

  >>> from rpy_options import set_options
  >>> set_options(RHOME='c:/progra~1/r/rw2011/')
  >>> from rpy import *

It looks like my version of R isn't supported by rpy. Does anybody know which version of R still works with rpy? I couldn't find anything about it. What about rpy2? Is it not supported by Galaxy?

Hi Sarah,

Although I haven't tested with 2.12.2, I know it works with at least 2.11.0. The error above is indicating that rpy is installed somewhere on python's path but the C-extension module for the version of R found on your $PATH (2.12.2) is not present. Most likely this is because rpy was already installed, and is being found earlier on python's path than the version of rpy that you installed. The older rpy was built against a different version of R.

If you find the original installation of rpy you should be able to determine from _rpy<version>.so what version of R it was built against, and put that version of R ahead of 2.12.2 in your $PATH.

Alternatively, you can find and uninstall the old version of rpy, which should allow your newer version to be found and used.

python's path can be determined with:

python -c 'import sys; print "\n".join( sys.path )'

There is some effort underway in the community to rewrite rpy tools to use rpy2, but this work is not yet complete.

Hope this helps, --nate

Thanks very much in advance for your help!

Best regards, Sarah Diehl _______________________________________________ To manage your subscriptions to this and other Galaxy lists, please use the interface at:

http://lists.bx.psu.edu/

Hi Galaxy Devs,

I think I found a bug in the handling of data_column parameters in combination with the data parameter they depend on. Unfortunately in some cases the popup menu to select a column number won't get populated. (I did a hg pull yesterday, but that didn't solve the problem.)

Here's what I have to reproduce the problem:

1. Make a new histroy 2. Upload a tabular file 3. Upload a FASTA file 4. Select Text Manipulation -> Merge Columns tool 5. Tada, you cannot select a column from the popup menu...

I've traced the problem to detection of a suitable initial value (history item) for the data parameter on which the data_column parameter depends. With the FASTA file uploaded last it will be selected as the initial value. It's not a tabular file, but it can be automagically converted to tabular (is_convertable( data ) == true). It won't show up though in the popup menu to select a history item for the data parameter. Apparently when the HTML form is generated the FASTA file is skipped and the tabular file is the only one in the popup menu.

In galaxy/lib/galaxy/tools/parameters/basic.py:

class DataToolParameter( ToolParameter ): ... def get_initial_value( self, trans, context ): """ NOTE: This is wasteful since dynamic options and dataset collection happens twice (here and when generating HTML). """ ... # The line below is number 1366 (in my version of this file.) # This may have bad effects for data_column parameters that depend on this data parameter; removing the 'or is_convertable( data )' will 'fix' it, # but is no good solution as it kills the ability to select data that can be converted automagically to a compatible data type. # The note above may explain why the convertable history item does not show up in the popup menu... #if data.visible and not data.deleted and data.state not in [data.states.ERROR, data.states.DISCARDED] and ( isinstance( data.datatype, self.formats) or is_convertable( data ) ): if data.visible and not data.deleted and data.state not in [data.states.ERROR, data.states.DISCARDED] and isinstance( data.datatype, self.formats ):

When the data_column parameter is initialized however class ColumnListParameter( SelectToolParameter ) -> def get_column_list( self, trans, other_values ) sees the FASTA file and hence there will be no dataset.metadata.columns and the popup menu remains empty...

If you have another tabular file in your history selecting that one and then switching back will update and populate the popup menu for the data_column parameter. If the order of the history items in the above example is reversed you also won't see the issue as a data set of the correct format is the "initial value".

Cheers,

Pi

------------------------------------------------------------------ Biomolecular Mass Spectrometry & Proteomics group Utrecht University

phone: +31 6 143 66 783 email: pieter.neerincx@gmail.com skype: pieter.online

visiting address: H.R. Kruyt building // room O607 Padualaan 8 // 3584 CH Utrecht // The Netherlands

mail address: P.O. box 80.082 // 3508 TB Utrecht // The Netherlands ------------------------------------------------------------------