R Add-on Package: qlapply
Kevin Ding (Kevin.Ding@anu.edu.au)
Current release version: 0.21 (Feb 20, 2003)
This
project has been done in co-operation with Markus Hegland, Ole Nielson
at Mathematical Sciences Institute and John Maindonald at Centre for
Bioinformation Science, the Australian National Universtiy.
Introduction
The qlapply system provides a parallel processing interface
for the lapply function of the open-source statistical
language/package R.
It has two components:
- An add-on package for R
- A client/server system which is written in Python
For qlapply to work on your system, Pyro (Python Remote
Objects) and R must both be installed, both on the client machine and
on the remote server. The system uses rsync, with an ssh protocol,
to handle file transfer. Once setup is complete, the R qlapply
package provides the following functions:
- autoStart() to initiate the server program
- qlapply(), which is a version of lapply() that is adapted
for parallel processing
- autoKill(), to terminate the server processes.
Copyright and Disclaimer
This software is licensed as Open Source software under the GNU public licence
version 2 or (at your option) later. It is subject to change without
notice. The qlapply system comes as-is, without warranty and
in no event can authors be held liable for any problems
resulting from this software.
Required Software
- The system should run whereever Python is available. However, it
is recommended to use Python version 2.2.2 or later.
- The system uses the Pyro (Python Remote Objects) to implement
the communication between client and server. Make sure a compatible
version of Pyro installed on both client side and server side. The
system may not work well if two different versiona of Pyro are installed.
- The system uses rsync with ssh protocol to transfer
files. It is user's responsility to set up rsync in such a way
that password issues are addressed (see
Installation and Configuration section about how
to do it)
Documentation
You can download
the qlapply manual (PDF file) that comes
with the qlapply distribution. Here is the online copy:
Download
Here you can download the source for qlapply:
Authors
- Kevin Ding (client/server system implementation, testing and
documentation, etc.)
- Markus Hegland (R source code, testing and various technical
advices)
- Ole Nielsen (assistance and advice with various technical
issues regarding python and Pyro)
- John Maindonald (testing, assistance with various technical
advices associated with R and reviewing the documentation)
Thanks
- Margaret Kahn for suggestions for system development.
- Edward C. Lang for testing on the CBiS cluster and
technical suggestions regarding ssh tunnel
- Matthew Wakefield for advice with ssh tunnel.
- ...everybody I forgot and all people that are interested in this
project. Thank you!
Return to TOP