Data manager in computational material science
 |
Introduction
Tilde (wwwtilda, ~) is a data manager in cheminformatics and computational material science. The nowadays situation in these disciplines with a plenty of incompatible data formats is rather confusing. Furthermore a new approach to the scientific data verification and exchange is desired because the traditional peer-reviewed journals and their supplementary materials sections do not solve this task properly.
Therefore Tilde aims to simplify data processing and exchanging for the scientists, dealing with popular computer simulation packages in material science (e.g. VASP, CRYSTAL, WIEN2K etc). Now the text outputs of these programs can be collected, and the calculated results therein can be visualized. The example focus is on phonon properties of the crystalline materials: vibrational spectra, band structures, densities of states, vibrational patterns visualization etc.
Moreover Tilde acts as an infrastructure for user's own scripts and applications, organizing them under single hood and assisting to pipe data throughout them. The third-party scripts and programs so could be easily integrated in Tilde with its application programming interface (API).
Technical details
Tilde is a client-server application. Usually for a client side of such applications one needs a web-browser as a user interface, while a server side resides somewhere remotely in the cloud.
Tilde however lives totally at the user own computer acting as a small local web-server. All user private data are therefore stored securely on a local computer. Tilde web-server generates a website with a snapshot of user calculations and modelling results. This website is protected by operating system and only available to current user, though he or she is able to share some of the results with others (see below). Tilde is written in Python programming language and therefore highly portable, the same code is used on board of Windows, Linux and Apple machines.
Pic. 1. Software architecture (red are supported programming languages, product names and tech standards are underlined)
Tilde is a networking application meaning the different running copies of it can communicate with each other located on different computers over the Internet. So its participants can connect and exchange data, compare their calculations and find out bugs together in modelling programs that they use for own data-mining.
Tilde was inspired by torrents, Skype and jabber applications. Their common feature is a peer-to-peer distributed syndication. So Tilde tries to adopt this idea to a data exchange in a scope of scientific problems, namely, computational modelling in physics and chemistry.
Requirements
The server side of Tilde is written in Python which generally speaking means it is able to run everywhere, on Windows, Unix, MacOS, Symbian computers and others where Python interpreter is implemented. It actually has been tested only on Windows 7, Windows XP and UNIX Debian though.
The client side of Tilde runs in any modern HTML5 web-browser (normally, Google Chrome, Mozilla Firefox or Safari). Flash and Java plugins are not required.
Licensing
Tilde has MIT-license. This means everybody is welcomed to use it for own needs or modify and adopt its source code.
Demo
Demo is actually running on a virtual Linux machine, which was kindly provided by Peterhof Telecommunication Center of St.Petersburg State University
Download
Tilde is a young actively developing project. API and networking features are not completed until now, so the sophisticated users are advised to wait for a stable release. For an early development code summary please refer to the downloads section. Have a look at the flyer therein as well:
Relative projects
Not only Tilde represents the need for a new instrument in scientific data management, manipulation and exchange. For example, other projects are:
- Scipio (java), https://scipio.iciq.es
- CMR (python), https://wiki.fysik.dtu.dk/cmr
- J-ICE (java), http://j-ice.sourceforge.net/
- CDF (python), http://kitchingroup.cheme.cmu.edu/cdf
- Quixote, http://quixote.wikispot.org/
- ESTEST (XQuery, python), http://estest.ucdavis.edu/
Not a secret that the world wide web itself was initiated as a mean of data exchange in the scientific community. The concept of a hyperlink came from scientific publications, being a way to reference the output of other scientists in the field. In this sense the principle of open data, open source scientific code and open standards is crucial. It is declared by an initiative group with a symbolic name Blue Obelisk (http://www.jcheminf.com/content/3/1/37).
Comments