SEMS

Outro -a.k.a- Where can I find $Stuff!?

SEMS — Tue, 23 Jan 2018 14:53:20 +0000

Ok guys, $Stuff is now distributed around different servers, domains, services, people… Here are some links for external users to find the resources:

SEMS web page

The wordpress site has been serialised to static web content and is now available through sems.bio.informatik.uni-rostock.de. All sems.uni-rostock.de/.* links should send proper HTTP redirects, so your browser will automatically guide you to the new location.

Trac

Everything trac-related is now available through GitHub. There are plenty of repositories at github.com/SemsProject, the trac-wikis and tickets are converted to GitHub pages at semsproject.github.io/.
The webserver should be configured to properly redirect individual projects, wikis, tickets etc… Thus, bookmarks and scripts should be working fine.

Compilation Products

During SEMS’ life time we created quite some binaries. People and pipelines are relying on their availability. However, they also needed to move and are now available here:

Maven artifacts moved to mvn.bio.informatik.uni-rostock.de (used to be mvn.sems.uni-rostock.de)
Binaries moved to bin.bio.informatik.uni-rostock.de (used to be bin.sems.uni-rostock.de)
JavaDocs moved to doc.bio.informatik.uni-rostock.de (used to be jdoc.sems.uni-rostock.de)

Again, webservers are configured to redirect people and pipelines automagically!

Maven

We continue developing our tools and we will provide them through Maven. As described above, our Maven repository has moved to mvn.bio.informatik.uni-rostock.de.
But even better: we’re now distributing our tools through Maven central! No need to integrate a third party repository anymore :)
However, we switched the group id. It is now de.uni-rostock.sbi – so update your POMs. But don’t worry, you don’t need to update all the package names, the classes are still packaged as de.unirostock.sems.*. Find our artifacts published under de.uni-rostock.sbi at Maven Central.
Maven central will only contain new artifacts. The repository at mvn.bio.informatik.uni-rostock.de will still deliver previous artifacts, and mirror Maven central for the new artifacts. Thus, feel free to keep our third party repository integrated in your workflows.

Tools

Online tools, demos and prototypes have moved as well:

The BiVeS web application is now running at https://bives.bio.informatik.uni-rostock.de. We installed a Proxy at the old URL (bives.sems.uni-rostock.de) that will invisibly hand over incoming tasks, so whoever integrated BiVeS does not need to worry. You should still switch to the new URL, just to make sure… ;-)
The CombineArchive Web interface already moved to https://cat.bio.informatik.uni-rostock.de/ a year ago. Projects, cookies, and files have been migrated and the webserver properly redirects traffic.
The CaRo interface is now available at https://caro.bio.informatik.uni-rostock.de/. A proxy is forwarding jobs, similar to the BiVeS web application above.
BudHat moved to https://budhat.bio.informatik.uni-rostock.de/. The dev version is available at https://budhat-dev.bio.informatik.uni-rostock.de/. Redirects are installed. However, there are already many real-life tools that showcase the power of BiVeS, so you BudHat only survived because of paper-links etc..
COMODI moved to https://comodi.bio.informatik.uni-rostock.de/latest/ already some time ago.
MoSt is now available at https://most.bio.informatik.uni-rostock.de.
M2CAT is still an issue, as it strongly relies on Masymos and Morre. However, there is no documentation on how to properly setup those tools :(
As soon as I find time to reverse engineer the stuff, M2CAT will properly be available at https://m2cat.bio.informatik.uni-rostock.de/.

By the way, did you recognize that we also deployed proper SSL certificates for all services? :)

Cool, but how…

..did I just do it? Maybe it’s interesting for one or the other, or maybe you need to archive/migrate your $stuff as well? Therefore a quick overview on how I’m handling different things.

WordPress

This website used to be a WordPress instance. It has been serialised to static HTML pages using HTTrack and a bit manual fine-tuning – see also Archiving a Website.

Services that can handle HTTP redirects

Services that just handle default HTTP traffic now reply with a permanent redirect (HTTP/1.1 301 Moved Permanently). That’s realised using tiny Nginx vhosts. For example, bin.sems.uni-rostock.de can now be found at https://bin.bio.informatik.uni-rostock.de/, and the following definition makes sure people are forwarded to the proper location:

1
2
3
4
5

server {
server_name bin.sems.uni-rostock.de;
listen 80 ;
rewrite ^/(.*)$ https://bin.bio.informatik.uni-rostock.de/$1 permanent;
}

Thus, a request on http://bin.sems.uni-rostock.de/CombineArchiveWeb/CombineArchiveWeb-0.4.9.war will be forwarded to https://bin.bio.informatik.uni-rostock.de/CombineArchiveWeb/CombineArchiveWeb-0.4.9.war. Try it yourself:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17

usr@srv $ curl -LI http://bin.sems.uni-rostock.de/CombineArchiveWeb/CombineArchiveWeb-0.4.9.war
HTTP/1.1 301 Moved Permanently
Server: nginx/1.13.8
Date: Wed, 24 Jan 2018 12:40:44 GMT
Content-Type: text/html
Content-Length: 185
Connection: keep-alive
Location: https://bin.bio.informatik.uni-rostock.de/CombineArchiveWeb/CombineArchiveWeb-0.4.9.war

HTTP/2 200
server: nginx/1.13.8
date: Wed, 24 Jan 2018 12:40:44 GMT
content-length: 47174359
last-modified: Thu, 08 Oct 2015 16:01:12 GMT
etag: "2cfd2d7-52199f8defcec"
accept-ranges: bytes
strict-transport-security: max-age=31536000

Services that can’t handle 301

The above is probably the preferred approach, as people/browsers/search-engines won’t recognize anything and the old URLs will smoothly disappear (they won’t be copied anymore etc).
However, not every tool/pipeline is able to work with 301 redirects… For example, the BiVeS web application accepts POST requests. Tools that send a POST to bives.sems.uni-rostock.de may also receive a 301 that BiVeS moved to bives.bio.informatik.uni-rostock.de, and they may look at the new URL if you’re lucky, but they (most certainly) won’t resent the POST data to the new location. Thus, all tools that integrate BiVeS will not work anymore… :(
Thus, I’m not responding with a 301, but the request will just be proxied to the new location. For example, the Nginx configuration for BiVeS looks like this:

1
2
3
4
5
6
7
8

server {
server_name bives.sems.uni-rostock.de;
listen 80 ;
location / {
proxy_pass https://bives.bio.informatik.uni-rostock.de;
}

}

This, however, is a bit sub-optimal as users won’t realised that BiVeS has moved…
That means I need to make sure this URL stays up and running!

Source code and wikis and tickets and stuff

Our git and trac stuff was migrated to GitHub (https://github.com/SemsProject/) and GitHub Pages (https://semsproject.github.io/). That has been made possible by Martin Peters, who developed a https://github.com/FreakyBytes/TracMigrator.

Monitoring

I’m a big fan of monitoring. Thus I started the migration with monitoring of web content using check_links. For example, I wrote a check to see if the page behind https://sems.bio.informatik.uni-rostock.de/projects/bives/ still contains the phrase highlighted reaction network. So a went around the many different sites and tools to define almost 100 different checks. This way, my monitoring solution would notify me as soon as a page/tool is not properly migrated and I can be sure that the HTTP redirects for all the URLs forward to the correct final locations (also with proper content negotiation for COMODI etc).
These checks are now properly integrated in my monitoring infrastructure, so that I can continuously be sure that links (in papers for example) stay valid.

Help from the whale

In the process of migration I encapsulated almost everything in Docker containers. This way it’s easier to handle the (legacy) content and a future migration will hopefully be much quicker.

But where is…?

Did we forget something? Is some tool not working properly anymore?
If you experience any problems please drop me an email.

The SEMS project came to an end

SEMS — Thu, 19 Oct 2017 07:21:12 +0000

As you may have guessed from this website’s quiet news section: The SEMS project came to an end. Tom was bravely holding the fort, spending the last few cents of our budget in July 2017. But do not worry, not all is lost!

SEMS peeps will continue their fights for improved management of computational models on various fronts:

Dagmar temporarily tries the real life outside academia.
Fabienne continues to do her PhD in Business Informatics.
Mariam remains in Rostock, but moved on to revolutionise machine learning.
Vasundra moved to Trondheim to work on causal statements and ontologies.
Martin and Tom managed to land a new project at the SBI in Rostock to continue work on the versioning front in collaboration with groups from Munich and Heidelberg.
Ron works towards finishing his PhD in a denbi project in cooperation with Heidelberg.

To keep our results and project code available in the future, we are moving most of our stuff to Github: semsproject.github.io/. Martin will try to keep most of the tools, prototypes, and demos alive, but probably under different domains. For example, the BiVeS web application already moved to bives.bio.informatik.uni-rostock.de and the Combine Archive Web tool is available from cat.bio.informatik.uni-rostock.de. We try to properly redirect browsers and tools to the new locations, but let us know if anything goes wrong.

If you have doubts/questions/miss us do not hesitate to drop SEMS people an e-mail. We’ll be more than happy to hear from you.

Also check our Outro.

SEMS extended until July 2017

SEMS — Mon, 30 Jan 2017 12:11:34 +0000

We are happy to announce that the SEMS project has been extended, and now runs until July 2017. During the remaining months, we will focus on improving our MOST tool for exploration of differences between versions of a model; and we will investigate how to measure differences between SED-ML files.

End of 2016, end of SEMS?

SEMS — Mon, 26 Dec 2016 18:04:47 +0000

It has been an exciting year for us, with good old and brand new collaborations, exciting projects, research stays close by and far away…

We are right in the middle of the 5th year of SEMS funding, meaning that the project will come to an end soon – we’re closing down on the 28th of February, 2017.

However – all our code is (or will be) available openly on github, bitbucket… And we are looking forward to what the future will bring :-)

New paper on exploration of SBGN maps using graph databases

vasundra — Mon, 05 Dec 2016 16:02:33 +0000

Vasundra Touré et al. published a paper on the STON framework in BMC Bioinformatics!
STON is meant to explore biological networks in SBGN using a graph database called Neo4j. The framework enables the query of SBGN maps.

Paper abstract:
Background
When modeling in Systems Biology and Systems Medicine, the data is often extensive, complex and heterogeneous. Graphs are a natural way of representing biological networks. Graph databases enable efficient storage and processing of the encoded biological relationships. They furthermore support queries on the structure of biological networks.

Results
We present the Java-based framework STON (SBGN TO Neo4j). STON imports and translates metabolic, signalling and gene regulatory pathways represented in the Systems Biology Graphical Notation into a graph-oriented format compatible with the Neo4j graph database.

Conclusion
STON exploits the power of graph databases to store and query complex biological pathways. This advances the possibility of: i) identifying subnetworks in a given pathway; ii) linking networks across different levels of granularity to address difficulties related to incomplete knowledge representation at single level; and iii) identifying common patterns between pathways in the database.

Read more at: http://www.biomedcentral.com/1471-2105/17/494.
The framework resources are available in sourceforge.

New paper on standards and software for whole-cell modeling

SEMS — Tue, 19 Jul 2016 07:01:34 +0000

After the successful Whole Cell Modeling Summer School in Rostock in March 2015, we ran a second meeting in Salt Lake City in 2015, co-located with COMBINE 2015. Both were practical meetings for young scientists aiming at recoding a Whole Cell model in standard formats and using open tools.

We provide all code on github. The outcomes of the meetings, a discussion on the suitability of standards for whole cell modells, and our experiences with organising hands-on educational events are described in our new paper entitled: “Toward community standards and software for whole-cell modeling” in the journal IEEE Transactions on Biomedical Engineering.

Paper abstract:
“Whole-cell (WC) modeling is a promising tool for biological research, bioengineering, and medicine. However, substantial work remains to create accurate, comprehensive models of complex cells.
Methods: We organized the 2015 Whole-Cell Modeling Summer School to teach WC modeling and evaluate the need for new WC modeling standards and software by recoding a recently published WC model in SBML.
Results: Our analysis revealed several challenges to representing WC models using the current standards.
Conclusion: We, therefore, propose several new WC modeling standards, software, and databases.
Significance: We anticipate that these new standards and software will enable more comprehensive models.”

Open access paper

A container for the CombineArchive Web interface

SEMS — Fri, 10 Jun 2016 07:35:59 +0000

Following the BiVeS-WebApp example, the CombineArchive Web interface jumps onto the Docker hype as well! It is now also available as a container from the Docker hub and we will automatically create new versions of the image with every release of the application.

To run the application just call the following command:

1	docker run -it --rm -p 1234:8080 binfalse/webcat:latest

This will bind your host’s port 1234 to the interface of the CombineArchive Web application in the container and you’ll be able to access the web interface at http://localhost:1234. There are a few options to run and configure the container, just have a look at the binfalse/webcat Docker Hub repository.

Thus, setting up the web interface was never easier nor cleaner. Plus: You will increase the privacy of your users and may get better performance from your web servers. :)

Let us know if you have problems with the setup.

BiVeS-WebApp as a Docker image

SEMS — Tue, 31 May 2016 21:07:46 +0000

Today we compiled the BiVeS web application into a Docker image. Using Docker it is very simple to run the application on your desktop, your server, your cloud…

The Docker image is available from the Docker Hub as binfalse/bives-webapp. To pull the latest version of the BiVeS web application just call

1	docker pull binfalse/bives-webapp

Running the tool is almost as simple pulling it:

1	docker run -it --rm -p 1234:8080 binfalse/bives-webapp

This call will bind the container’s web server to your machine’s port 1234. Thus, if you open your web browser and go to http://localhost:1234 you will already start talking to BiVeS!
To verify that BiVeS is working correctly, you can try to compare two files using the following curl command:

1
2
3
4
5
6
7
8
9
10
11
12
13

curl -d '{
"files":
[
"http://budhat.sems.uni-rostock.de/download?downloadModel=24",
"http://budhat.sems.uni-rostock.de/download?downloadModel=25"
],
"commands":
[
"SBML",
"reactionsDot",
"reportHtml"
]
}' http://localhost:1234

Replace http://localhost:1234 with the settings of your environment if you didn’t run the above command lines on your local machine.

We updated our maven configuration to automagically create and publish new docker image with every release of the BiVeS web application.

New paper on Functional Curation

SEMS — Sat, 13 Feb 2016 14:47:39 +0000

Martin Scharm’s paper about the Cardiac Electrophysiology Web Lab, work done during his research stay at the University of Oxford, is published in the current issue of the Biophysical Journal! Hooray :)

We have created an online resource for the characterization and comparison of electrophysiological cell models in a wide range of experimental scenarios. The details of the mathematical model (quantitative assumptions and hypotheses formulated as ordinary differential equations) are separated from the experimental protocol being simulated. Each model and protocol is then encoded in computer-readable formats. A simulation tool runs virtual experiments on models encoded in CellML, and a website (https://chaste.cs.ox.ac.uk/WebLab) provides a friendly interface, allowing users to store and compare results.

Martin Peters visits the JJJ lab in Stellenbosch

Martin Peters — Wed, 03 Feb 2016 14:45:27 +0000

Martin Peters is at the Stellenbosch University, South Africa to visit the research group of Jacky Snoep. His goal is to facilitate closer collaboration with the JJJ lab and work on a possible repository for SED-ML simulation descriptions together with the development team of JWS Online. Martin will be at the JJJ lab until June 2016.