Functionality Plugins ~ console plugin python
The first release of the Gephi’s Python Console plugin is finally available for download. Through this plugin, you can execute queries and manipulate the graph structure by typing commands on a scripting console, making it a very powerful and concise tool to work with.
This console started as a joint proposal with the GUESS project which aimed at porting the Gython language as a console plugin for Gephi during the Google Summer of Code 2011 program. This project was mentored by Eytan Adar, from the GUESS project, and co-mentored by Mathieu Bastian, from Gephi.
After installation, the plugin can be accessed through the left slide on Gephi’s UI or through the Window > Console.
Among the features available through the console scripting language are:
- Inspired by the Gython language from the GUESS project (i.e. additional operators for manipulating graphs).
- Based on Jython 2.5.2, which implements the Python 2.5 specification.
- Graph structure and attributes manipulation.
- Filtering support.
- Support for running layouts.
- Export API support (i.e. generate PDF/GEXF/PNG exports from within the console).
- Batch scripts support for automating tasks.
- Loads third-party Python libraries.
For example, if you would like to filter the male persons from your social graph and make the result visible, you could simply type the following command:
visible = g.filter(gender == "male")
If you want set this subgraph nodes’ color to blue, you could just type:
g.filter(gender == "male").nodes.color = blue
Or, even better, if you want to color both the nodes and edges of this same subgraph, just type:
(gender == "male").color = blue
Implementing batch scripts for automating tasks is very easy. Just save the script as a file with the py extension and load it with the execfile function. For instance, a batch script for creating a random graph with 50 nodes would look like this:
And to run the script, you would call the following command on the console:
>>> execfile("C:\\Users\\user\\Documents\\script.py")
Installation
The Scripting Plugin can be installed through the Gephi’s Tools/Plug-ins menu:
- Go to the Tools > Plug-ins menu on Gephi’s user interface.
- Open the Available Plugins tab.
- Click the “Reload Catalog” button, to make sure you have an up-to-date catalog.
- Select the “Scripting Plugin” on the list and click the Install button.
After installation, the plugin can be accessed through the left slide on Gephi’s UI (as shown in the image below) or through the Window > Console menu.
Documentation
The console’s documentation at the wiki has instructions for downloading and installing the plugin.
Feedback and Contributing
This plugin has been developed by Luiz Ribeiro since the Google Summer of Code 2011 program and we’re really looking forward to see what you’ll be able to do with this new tool.
The plugin’s source code is hosted on GitHub and is open to your contributions. If you find any bugs or have requests or feature ideas make sure to open a new ticket on the Issues Tracker.
Functionality ~ gsoc index performance
My name is Ernesto Aneiros and during this Google Summer of Code I am working on the Attributes Disk Store.
The problem
In Gephi, Attributes are the data that is associated with nodes and edges. As graphs grow larger and larger, attributes occupy more memory even though many times they are not essential to the end-user when he is only applying transformations or algorithms to the graphs. These attributes can be of different types, from simple Java primitives (byte, char, int, String, etc) to Gephi’s internal data types (lists of primitives or versioned data). The idea for the project was to have a combined memory/disk cache system to partially off-load these attributes to disk. The system should have a well-designed cache system to handle heavy read access on the most-accessed elements.
The Solution (1st iteration)
Lucene is one of the most popular text searching engines and a flagship Java open source project. Lucene is capable of handling and indexing millions of records while remaining performant, and when the idea for the Attributes API cache was born, Lucene was first considered for the role of the data store, and as added bonus Gephi will get full-text search capabilities with almost no extra effort. When analyzing the problem, the following criteria were developed to judge a possible data store:
- Reliable (resistant to corrupted disk data, failed transactions, unexpected errors)
- Fast
- Transparent (minimum complexity exposed to the end-user)
While Lucene complies with items 2 and 3, the approach when dealing with corrupted indices in Lucene is to rebuild from scratch therefore failing item 1. This doesn’t pose a problem to Lucene because in the context where it is supposed to be used (indexing of external information), input data is always available separately from the index and can be accessed if needed. In Gephi, however, this is not the case. Once Attributes are loaded from disk they remain in memory until saved back to file. If an error occurs during a disk store transaction the end-user can end up losing a day’s work, certainly not acceptable.
The Solution (2nd iteration)
After Lucene was ruled out as a contender for a data store, several options were considered, including using embedded SQL databases and using a combination of Ehcache plus BerkeleyDb. Both options bring a lot to the table and embedded databases in Java have achieved impressive results in performance when compared to other mainstream database systems (see projects H2 and HSQL for example). Ehcache + BerkeleyDb however win when complexity is considered since they introduce almost no translation layers between Gephi and the cache. Both solutions are good fits for the problem but in the end the balance tilted in favor of Ehcache + BDB because the complexity consideration.
Optimizing Ehcache and BerkeleyDB
Even though Ehcache provides a great deal of functionality and features, it was relatively easy getting up to speed with it. The documentation provided online was very complete with code samples available and detailed explanations. In almost no time an in-memory cache was up, running and being tested. Traditionally cache sizes have been specified as the amount of max elements that they can hold. In the 2.5 BETA of Ehcache a new feature was introduced that allowed sizing the caches by memory consumed instead of elements held. For our project this is a killer feature since we can now expose a single option to the user, letting him specify how much memory the cache should consume. Even though using the new feature proved a little more complicated than expected we obtained great feedback from the Ehcache community, specially from alexsnaps and Mike Allen, which helped us to solve the issues we were having.
BerkeleyDB on the other hand, is a very complex piece of software. With years of development under the belt, BDB has evolved to be a very robust and flexible database. In fact, it is so flexible that can be used as full blown database supporting queries, a simple key/value datastore or with a front-end that exposes a Java collections map that greatly simplifies its use. All of this flexibility does not come free though, configuring and optimizing BerkeleyDB requires delving into details about transactions, buffers, log file sizes and BDB internals. However the tools are there and the information provided is quite good, especially the FAQ and the optimization section.
Integration with Gephi
Since ease of use and transparency are important considerations for the end-user of Gephi, only the minimal configuration options are exposed in the preference panel of the disk cache, but an Advanced tab provides more control for those who want it.
The General settings tab, where cache can be enabled or disabled and the memory usage configured.
The advanced settings tab allows a more advanced user to configure several of BerkeleyDB’s options.
The Disk Store in Action
Memory consumption without the disk store. It reaches 400MB.
Memory consumption with the store, after load it drops below 400 MB. Note how load time increased due to disk operations, a trade-off to consider when using the store.
Known issues
The project is still in development. Being memory saving the main goal of the disk store project, results are not good enough yet because of several reasons.
While BerkeleyDB provides a very convenient way of storing bytes in disk, it is still a database oriented software and therefore it is not the most suitable solution for out project because of large memory usage to caching data, building and maintaining its index (features desirable for databases but not for this project).
Trying to reduce BerkeleyDB memory usage with its settings will produce quite different results in different systems or even in the same system. The benchmark above shows not bad results but it is not always the case. A better control of maximum heap growth can be observed but still with memory usage peaks that prevent better saving.
The conclusion is that it is a priority to replace BerkeleyDB with other disk persistence system or create one specifically designed for Gephi disk store.
It is also known that graphs with more complex data like strings or lists will always benefit more from a disk storage system than graphs with simple data like integers or booleans. An idea is to always store simple data in memory because indexing in in the disk is going to need as much memory anyway, or even more.
On the other hand, Gephi works and was designed with in-memory data structures in mind. Adding a cache/disk store to the system is bound to create integration issues with other parts of the codebase. For example the GEXF file importer tends to load large portions of the graph file to memory while parsing it, which is not so good in memory constrained environments and using the cache here will not make a difference. One of these issues is regarding the handling of data in files with .gephi format. Due to the way that .gephi files are imported, some integration problems still need to be debugged in the disk store to work properly.
Looking to the future
This GSOC project is only scratching the surface of what a memory + disk cache system can achieve. In the future BerkeleyDB could be replaced with other persistence provider, and it doesn’t necessarily has to persist locally to the disk. For example replacing BerkeleyDB with a datastore like Cassandra, or maybe some RDBMS.
Conclusion
While the Data Store API introduced by this project is still taking its first steps and can be significantly evolved, it has helped ironing out many issues and has paved the way for bigger and better improvements. Working during this summer has been a great experience and I have been able to share with great mentors like Eduardo Ramos, who knows the Gephi codebase in and out. I hope the work of all of Gephi’s GSOC’ers becomes the starting point for many new features and enhancements that the community will surely appreciate. Happy coding and see you next summer!
Design Functionality ~ dynamics gsoc timeline
My name is Daniel Bernardes and during this Google Summer of Code I am working on the new Timeline interface.
Dynamic graphs have been the subject of increasing interest, given their potential as a theoretical model and their promising applications. Following this trend, Gephi has incorporated tools to study dynamic networks. From a visualization perspective, a critical tool is the Timeline component, which allows users to select pertinent time intervals and display and explore the corresponding graph. The challenge concerning the timeline was twofold: redesign the component to improve user experience and add extra features and introduce an animation scheme with the possibility to export the resulting video.
Together with my mentors Cezary Bartosiak and Sébastien Heymann, we have proposed a new design for the timeline component featuring a sparkline chart in the background of the interval selection drawer (which is semi transparent): this feature will help the user to focus on particular moments of the evolution of the dynamic graph, like bursts of connections or changes in graph density or other simple graph metrics. Current metrics are the evolution of the number of nodes, the number of edges and the graph density. The sparkline chart was preferred to other chart solutions because it does not add too much visual pollution to the component and adds to the qualitative analysis. The interaction with the drawer remains globally the same of the old timeline, to guarantee a smooth transition for the user.
To implement this feature we have used the chart library JFreeChart (a library already incorporated to Gephi), customizing their XYPlot into a Sparkline chart by modifying their visual attributes. To display the Sparkine, one needs to measure the properties of the graph in several time instants of the global time frame where the dynamic graph exists. This represented a major challenge, since the original architecture did not allow the timeline component to access (and measure) the graph in particular instants of time; the solution was to introduce a slight modification to the DynamicGraph API to provide an object which gave us snapshots of the graph at given instants. Other challenges we dealt with included the automatic selection/switching of real number/time units in the timeline (depending on the nature of the graph in question) and sampling granularity of the timeline.
Another breakthrough of this project was the introduction of the timeline animation. Once the user has selected a time frame with the drawer it can make it slide as the corresponding graph is being displayed on the screen. Besides the technical aspects of interaction between the timeline and the animation controller, there were also an effort to calibrate the animation (ie, in terms of speed and frames) so it would be comfortable and meaningful for the user.
As far as the UI is concerned, the component has gained a new “Reset” button next to the play button which activates the timeline drawer and displays the chart. It also serves to reset the drawer selection to the full interval when the timeline is active. The play button gained its original function, that is, to control the animation of the timeline — instead of activating the selection.
Finally, the animation export to a video format revealed to be more tricky than expected and couldn’t be finished as planned. There were several setbacks to this feature, beginning with the selection of a convenient library to write de movie container: it turns out that the de facto options available are not fully Java-based and need an encoder working in the background. The best alternative I found was Xuggler, which is based on ffmpeg. Also, obtaining screen captures of the graph to were a little bit tricky so I have exported SVG images from the graph corresponding to each frame, converted them to jpeg and than encoded them though Xuggler to a video format. As one might expect, this solution is not very efficient in terms of time, so Mathieu Bastien and my mentors suggested me to wait for the new features from the new Visualization API that would make this process simpler.
In addition to current bugfixes and minor improvements concerning the timeline and the animation, the movie export remains the the next big step to close this project. If you have questions or suggestion, please do not hesitate! The new timeline will be available in the next release of Gephi.
DB
Design Functionality ~ gsoc infovis preview
My name is Yudi Xue and during this Google Summer of Code am glad to work on the Core evolution of Gephi.
Current API in the Preview module provides too many granular methods and classes. Developers are clueless about how they may extend the component. In this project, we do not seek to expand what the Preview module already have to offer. Rather, we focus on making the Preview module easy to learn, easy to use and easy to extend to the Gephi developers. The new API will allow developers to focus on particular parts of the module. They may specify a new visual algorithms just by implementing a new type of Renderer, such as edge bundling and convex hull. They may also extend the RenderTarget to allow display or export visualization to different platform.
The user story
We took the infovis reference model into consideration when we started designing the new infrastructure. The infrastructure aims to provide support to a visualization-preview workflow:
raw data -> the data builder -> renderers -> render targets.
In particular, the raw data is the graph associated with the current gephi workspace. The data builder (DataBuilder) will interpret information associated with the nodes and edges and generate Item objects for Preview use. The Item objects are immutable objects that are either node item (NodeItem), edge item (EdgeItem) or item group (GroupItem) specified from the graph workspace or data lab. We append “Item” to refer that they are data rather than display objects. After the data has been imported, the preview controller (PreviewController) will associate each type of entity items with Renderer objects. Renderer objects are functional procedures that describe how an item should be drew. While we give information to an Renderer object what it is going to draw, we also tell it what RenderTarget it will use. By default, we provide ProcessingRenderTarget, PDFRenderTarget and SVGRenderTarget. All RenderTarget objects contribute to the RenderTarget API, which provide granular drawing functions that can be used by developers to form advanced visual algorithm. In addition to the workflow, we will provide a flexible properties structure to the Preview module so it may be used to provide listener to user interface commands. The property will allow dynamic dependency where grouped properties can listen for a single parent property.
The code below demonstrates how a Renderer to a particular Item type could be updated at runtime.
Code sample:
PreviewController prc = (PreviewController)Lookup.getDefault().lookupItem(
PreviewController.DEFAULT_IMPL).getInstance();
prc.loadGraph();
// Load graph from workspace
prc.updateRenderer(NodeItem.class, new Renderer() {
// How I want to draw a node, edge, or item types.
// Specify your procedural visualization algorithm here
@Override
public void render(Item item, RenderTarget rt) {
NodeItem ni = (NodeItem) item;
rt.drawImage(..);
rt.drawline(..);
}
// The RenderTarget will pick up the properties and draw the rest..
});
The big picture
Speaking of API flexibility, the Preview API goes from constrained to flexible in the direction from DataBuilder to RenderTarget. Here is the big picture:
Current progress
- done:
- a working copy based on the new architecture
- added ProcessingRengerTarget
- added GroupRenderer (Convex hull)
- added ImageRenderer
- basic unit testing
- basic functional testing against updating Renderer in Preview API at runtime
- in progress:
- Property support
- Selfloop, curved edge drawing
- PDF and SVG RenderTarget implementation
Here is a screenshot of the new system with convex-hull enabled:
Code practice
The code base is under active development at https://code.launchpad.net/~yudi-xue/gephi/gephi-preview. The code base includes the PreviewAPI module and the PreviewImpl module.
Lookup API
We make use of Netbeans Lookup API to instantiate singleton and use Lookup. Template to ensure the correct implementation been called.
For example, to call the default PreviewController constructor, we call:
/*
* DEFAULT_IMPL is defined in the interface.
* It refers to default implementation class
* "org.gephi.preview.PreviewControllerImpl"
*/
(PreviewController)Lookup.getDefault().lookupItem(PreviewController.DEFAULT_IMPL).getInstance();
Accordingly, you may choose to use the API with your implementation by creating a Template that points your implementation class.
Functional Tests
During the development, we are creating functional tests against our own API for the purpose of both flexibility and stability. the “PreviewAPIFunctionalTest”
Conclusions
Our goal is to bring modularity and extensibility to the Preview module. We aim to deliver the freedom in defining your own visual algorithms (Renderer) and user interaction (Property) and make use of API without thinking about the detailed mechanism. I would like to give my thanks to Dr. Christian Tominski, Mathieu Bastian and Sébastien Heymann for their support and feedback, which is critical during the development for the new architecture.
Functionality ~ developer gsoc maven toolkit
My name is Keheliya Gallaba and during this Google Summer of Code I am working on the Automated build system for Gephi. The goal of this project is to add Maven build support to Gephi and set up a continuous integration system to fasten the release process. The Netbeans Platform, which Gephi is built upon, natively uses Apache Ant to compile, build and package the application. But now there is also a variant of NetBeans which uses Apache Maven as the build system. There are several reasons that make moving into a Maven based system worthwhile.
Maven vs Ant
The existing Ant build system for building NetBeans Platform-based applications which is called Ant Build Harness is very intuitive, and needs almost no initial setup. The set of standard Ant scripts and tasks can be easily triggered by the IDE or by the command line. But there are reasons that Ant might not suite a rapidly growing, multi-module project like Gephi. The Gephi project consists of a team of developers who work on dependent modules and plugins. These modules have to be composed to the application regularly. With a large number of modules, with many small packages, and with multiple projects with many inter-dependencies and external dependencies, its essential to manage different versions and branches with their dependencies. And reusing modules with the Ant build harness is not that intuitive.
But Apache Maven is introduced as a standard, well defined build system that can be customized. It uses a construct known as a Project Object Model (POM) to describe the software project being built, its dependencies on other external modules and components, and the build order. It comes with pre-defined targets for performing certain well-defined tasks such as compilation of code and its packaging. It makes dependency management very easy and efficient with the concept of repositories. Most importantly in maven unique coordinates: groupId, artifactId, version, packaging, classifier identifies an artifact which can be uploaded or retrieved from a repository. This helps to easily build modules which depend on other modules.
Work completed so far
This project involves digging deeper in to the Gephi’s architecture and understanding dependencies, building and packaging. Gephi includes 100+ submodules categorized into Core, UI, Libraries and Plugins sections. NBM, which stands for “NetBeans module”, is the deployment format of modules in NetBeans. It is a ZIP archive, with the extension .nbm, containing the JARs in the module, and their configuration files. NBM files can be manually installed using the Update Center and choosing the option for installing manually downloaded modules, or they can be downloaded and installed directly from netbeans.org or another update server.
I’m happy to say that I was able to successfully mavenize 75 modules and continuing to complete the rest. I primarily used the NetBeans Module Maven Plugin for this, which now comes built in with NetBeans 6.9 and 7.0 IDEs. Currently NBM handles the tasks like defining the ‘nbm’ packaging by registering a new packaging type “nbm” so that any project with this packaging will be automatically turned into a netbeans module project, creating nbm artifacts and managing branding. It is also capable of populating the local maven repository with module jars and NBM files from a given NetBeans installation.
Some third party libraries used in Gephi are not maintained in any public Maven Repositories. So I had set up a local Sonatype Nexus Repository to store and serve these dependencies. Basic functionalities of a repository manager like Sonatype are:
- managing project dependencies,
- artifacts & metadata,
- proxying external repositories
- and deployment of packaged binaries and JARs to share those artifacts with other developers and end-users.
We are in the process of setting up a Sonatype Nexus Repository in official Gephi server as well, so not only these third party jars, but the Gephi releases such as the Gephi Toolkit can be served as a maven dependency to maven-based projects all over the world.
Challenges faced during the process
- Researching on existing large scale applications using NetBeans RCP and Maven
- Finding documentation on handling Netbeans specific ant tasks, now in Maven
- Managing transitive dependencies and versioning (specially with slight defferences of Maven and NetBeans difinitions)
- Compilation and Test Failures.
Continuous Integration
While Maven migration is going on, I also looked in to the other aspect of the project, setting up of a continuous integration server. Main benefits of such a system are:
- checking out source from source control,
- running clean build,
- deploying the artifacts in a repository
- and running unit tests.
Furthermore it can notify developers via Email, IM or IRC on Success, Failure, Error and Warning in a build or simply a Source Code Management Failure. What this means is that when a project gets updated during development, the continuous integration system will try to build the project and will notify the developers if it ran into any issues. This is very useful when working on a multi-module project with many developers, like Gephi since a developer may unintentionally, by accident break the build since they are working concurrently on code and they may have unique configurations to their development environment that isn’t shared by other developers. I looked at the options of Apache Continuum, Hudson and Jenkins (A fork of Hudson) considering the criteria, being open source, supporting Ant & Maven and better integration with Java based projects.
Hudson is an extensible Continuous Integration Server built by Sun Microsystem’s Kohsuke Kawaguchi. Since the design of Hudson includes well thought-out extension points, developers have written plugins to support all of the major version control systems and many different notifiers, and many other options to customize the build process for example the Amazon EC2 plugin to use the Amazon “cloud” as the build cluster.
Continuum is described as a fast, lightweight, and undemanding continuous integration system built by Apache Maven team. It is built on the Plexus component framework, and comes bundled with its own Jetty application server. Like Maven, it is built on the Plexus component framework, and comes bundled with its own Jetty application server. It uses Apache Derby, a 100% Java, fully embedded database for its persistence needs. All these reasons make Continuum self-reliant, and also particularly easy to install in almost any environment.
After considering all of these reasons I settled on Apache Continuum because of the ease of setting it up, configuration and out-of-the-box support for Bazaar. Bazaar is the distributed version control system used in Launchpad for managing the source code, when lot of developers work together on software projects like Gephi. I have set up a local instance of Apache Continuum to check out and build the ant-based Gephi hourly. In the future we can host this in the Gephi server to notify the developers and administrators.
Future Work
Since the initial foundation has been laid out, it will be quite convenient to complete the rest of the planned work. These will include completion of mavenizing rest of the modules, creating .zip distribution, properly running the final project being developed and setting up the infrastructure at the Gephi server.
I would like to thank my mentors Julian Bilcke, Mathieu Bastian and Sébastien Heymann for providing all the guidelines and support for making this project a success. You can find my ongoing work at this repository: https://code.launchpad.net/~keheliya-gallaba/Gephi/maven-build
References
Subscribe to your RSS feed
