We are delighted to be working with OMII-UK on the refactoring of OSCAR and OPSIN – our tools for chemical entity extraction and chemical name2structure conversion. There is now a lot of interest and uptake in these tools and we are now running a 1-day workshop. Here are the details – anyone is welcome to come – first-come first served.
As you may know we are working with OMII-UK under the JISC ENGAGE program to refactor, repackage and enhance the OpenSource OSCAR and OPSIN software. OSCAR is now ca 7 years old and is widely used in chemistry and bioscience for the identification of chemical entities in text. Informal studies have shown it probably has the highest precision and recall of any commonly used tool. OPSIN’s name2Structure has been informally tested against corpora of names, and again it is not far behind the leading commercial tool and has a smaller error rate.
We get feedback from many groups and have started to get offers of help in kind to enhance these tools. We are therefore inviting a number of collaborators and early adopters to a relatively informal workshop to explore what you would like to see done to OSCAR/OPSIN.
This is provisionally scheduled for 9th July 2009 and will be held in the Unilever Centre, Cambridge. There is no charge for attendance at the workshop and lunch will be provided. The programme will include presentations of the experience of 2-3 early adopters; a wish-list
session; a tour through the architecture; and a roadmap of the projected enhancements through the ENGAGE project. We would welcome input from you into the future of OSCAR/OPSIN, and how the evolution of these applications would best meet your future needs.Anyone interested in OSCAR/OPSIN, both non- and for-profit, is welcome to attend and should contact s.brewer@ecs.soton.ac.uk at OMII-UK.
For more information about OMII-UK, please see http://www.omii.ac.uk/, and for ENGAGE, please go to http://www.engage.ac.uk/. You can find out more information about OSCAR from http://oscar3-chem.sourceforge.net/, and http://wwmm.ch.cam.ac.uk/.
Peter, as you may know, I’m a fan of OPSIN in particular:
http://depth-first.com/articles/tag/opsin
I have some suggestions on how to increase contributions to the OPSIN code base.
Way back in 2006 I think, I helped factor out OPSIN into its own source tree:
http://sourceforge.net/project/showfiles.php?group_id=169991&package_id=207398
I did this because I wanted to have a lightweight nomenclature parser that could be included into other cheminformatics libraries.
I’ve recently had a look at the OSCAR/OPSIN SVN tree, and am puzzled as to why OSCAR and OPSIN are still intermingled as they are, especially given the apparent maturity of the OPSIN code base. OPSIN is useful on its own as a chemical nomenclature parser, and I found only light use of the OSCAR resources when looking through the source code.
Given that it’s entirely possible to factor OPSIN out into its own project, why not do so?
And if you’re looking for ways your team can easily incorporate new features, unit tests, documentation, and dictionary improvements created by others, I’d suggest hosting the project on GitHub and switching to git. The alternative is a painful process of creating and emailing patches, discussions about svn write access, and other frictional stuff that’s a powerful demotivator for potential contributors.
I know this is asking a lot and might not be feasible. Just some ideas.
BTW, I’d really like to be at the event, but a continent and a large body of water stand in the way. Any chance someone in your group has a digital camera and a knowledge of how to upload videos to YouTube or similar?
@Rich many thanks
I have some suggestions on how to increase contributions to the OPSIN code base.
Way back in 2006 I think, I helped factor out OPSIN into its own source tree:
http://sourceforge.net/project/showfiles.php?group_id=169991&package_id=207398
PMR: Thanks and we have to make sure that this contribution is recorded
I did this because I wanted to have a lightweight nomenclature parser that could be included into other cheminformatics libraries.
I’ve recently had a look at the OSCAR/OPSIN SVN tree, and am puzzled as to why OSCAR and OPSIN are still intermingled as they are, especially given the apparent maturity of the OPSIN code base. OPSIN is useful on its own as a chemical nomenclature parser, and I found only light use of the OSCAR resources when looking through the source code.
PMR: We’ve been looking at this in depth and the two are completely separable and they are being separated. That’s what OMII and daniel have agreed
And if you’re looking for ways your team can easily incorporate new features, unit tests, documentation, and dictionary improvements created by others, I’d suggest hosting the project on GitHub and switching to git. The alternative is a painful process of creating and emailing patches, discussions about svn write access, and other frictional stuff that’s a powerful demotivator for potential contributors.
PMR: Thanks for this and we’ll bounce this around – we are certainly not religiously wedded to SF
BTW, I’d really like to be at the event, but a continent and a large body of water stand in the way. Any chance someone in your group has a digital camera and a knowledge of how to upload videos to YouTube or similar?
Good idea…
Pingback: Unilever Centre for Molecular Informatics, Cambridge - Cambridge/OMII Workshop on OSCAR and OPSIN « Peter Corbett