from Ernst de Ridder <hnrid...@t-online.de> reply-to Ernst de Ridder <hnrid...@t-online.de> to nathann.co...@gmail.com cc Van Bang Le <l...@informatik.uni-rostock.de> date 4 March 2010 19:51 subject ISGCI <-> Sagemath interfacing
> I am a user and a developper of the Graph Library in the software Sage > Math ( http://sagemath.org/ ), an attempt to build a free alternative to > Maple... I would be very interested in using your database inside this > software, first as a handy way to question it, but also in order to > find, given, for example, a graph class, which kind of algorithm would > be best to solve on it a problem of maximal clique or vertex coloring. > > Could you tell me in which way it would be for us possible to use your database ? Is it licensed under some GPL-compatible license ? Would we need to directly send requests to you website ? Hi Nathann, thank you for your interest in ISGCI. In order to make the discussion on how to interface a bit easier, first a bit of background on how ISGCI is implemented. ISGCI started life as an applet that could draw relations between graph classes. Accordingly, most of the system is still written in java, although there are some auxiliary tools that are written in Perl and in Python. Originally, all there was, was the applet. Once we began to add information to the system, we started to store information in (generated) web pages. Currently, all the information on the classes is visible in the web pages and the applet can be use to generate coloured drawings of the relations between graph classes. The process goes roughly like this: input xml files -[java apps]-> derived xml files -[xslt scripts]-> web pages input xml files -[java apps]-> derived xml files -[java applet]-> drawing As you will probably have noticed, the applet doesn't work so very well anymore in modern java versions and the graph class pages look a bit dated as well. Currently, we're testing ISGCI v3, which will work in java 1.6 and have nicer looking web pages as well. Now, if you want to access data from ISGCI, I see three ways: 1. We can give you copies of the derived xml database files. Advantage: You're totally independent of ISGCI. Disadvantage: You're missing out any updates to ISGCI, you have to do your own parsing of these files. Below I attached snippets from these files. 2. You write some clever script to extract the data you want from the ISGCI web pages. Advantage: No work for us:-) Disadvantage: You have to rework your scripts whenever the format of the web pages change. 3. You extract the data you want from ISGCI using a cgi interface. Advantage: Flexible, always up-to-date data. Disadvantage: not independent of ISGCI. I'd say 2 is not a good idea, leaving 1 and 3. What I like about 3 is that we have been thinking about a convenient way for the user of querying the system rather than browsing it. With such an interface we could swat 2 flies at once: Better querying inside ISGCI and querying from outside ISGCI. However, this depends on - what information you want to extract from ISGCI. Classes? Relations? - in what format do you want to get it. Drawing? Text? Some xml structure? A link to the graph class page in ISGCI? So could you give examples of the queries you'd like to do and the type of answer you'd expect? And would for your purpose on-line or off-line access be better? Ernst P.S. Quite an impressive package, sage. -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
