Hello! great idea of color of this siyte!

iYlw8V A number of universities have awarded her honorary degrees, and she earned a prestigious job on the staff of Detroit congressman John Conyers. In 1988 Roxanne Brown noted: "Thirty-two years after she attracted international attention for sparking the Montgomery Bus Boycott, Mrs. Parks's ardent devotion to human rights still burns brightly, like a well-tended torch that ignites her spirit and calls her to service whenever she is needed."

Hey, there is what you need.

And some more..
,

Konnichiwa!
Check this out!
,

*

And some more..
*

Marhaba!
Check this out!
,

At last...
*

And some more..
*

Email to me about GraphPath can now be sent to adv-list-graphpath@langdale.com.au

Hi,

yes - you could just send the modified libadapt.py if you like. GraphPath has had a lot of in-house use and has evolved a lot since I posted the open source version. If there are any users (such as yourself) I will post the the current in-house version and update libadapt.py.

- Arnold

adv@langdale.com.au


QT - lucie wrote:
>


--
Arnold deVos
Langdale Consultants
adv@langdale.com.au
< replied-to message removed by QT >

Hello,

I am using the latest version of rdflib. I had to use the GraphPath product to search on the RDF graph I use more efficiently and it works pretty well.

But to have GraphPath working with the latest rdflib, I had to update a bit the file libadapt.py:
- TripleStore has been replaced by ConjunctiveGraph
- the method "triples" brings back (subj, p, o), cg instead of (subj, p, o) before.
Are you interested for me to send you a patch to update the product? If so, tell me where should i send it.
Lucie

Gods Blessings to you and your family. The Lord is Faithful and I continue to look forward to the moving of Gods Spirit in your lives. Visit estate norco real webpage devoted to estate norco real. norco window webpage devoted to norco window. and peace!

Re a new release of GraphPath:

I've been too busy using GraphPath in a data migration project for a client! And on vacation in Central Australia.

I trust that the present release is solid and usable. I don't plan to change the public API for writing and using expressions, although the adapter protocol (for mating it to rdflib, Redland or other) will change.

This week I plan to upload the parts where I have regression tests already. That would cover the following:

- support for writing

   Node(x.Bush)/Property(y.hasPolicies)[Property(approvedBy)/Node(x.Kerry)]

as
    x.Bush/y.hasPolicies[y.approvedBy/x.Kerry]

- proof: find all paths matching an expression and find all anticedents of an inferred statement

More test cases are needed before I can release forward chaining inference and detection/resolution of conflicting statements.

Regards,
Arnold

OK, I still haven't heard from you Arnold -- are you still there? BTW, how far away is this horizon of which you speak? Maybe I should wait ... or is it more like "somewhere over the rainbow"?? ;)

Hi Arnold,

Thanks for the quick reply! No, I don't have your address (at least the one I tried didn't get thru). Mine is stephen dot c dot waterbury at nasa dot gov. You can send yours to me at that address.

Cheers,
Steve

Re GraphPath source code format:

It is a good idea to normalise indents - I'll do that.

Not sure about epydoc annotations in comments. Probably worthwhile. I wonder if python's new decorators are good for putting the same metadata into the program itself? Anyway I will look at the latest epydoc information.

- Arnold

Hi Steve,

Thanks very much for the kind words about GraphPath. I am very interested in some of the areas you mention so send me an email (have you got my address). Two things on the horizon for me that are relevant:

First, I have a new wave of GraphPath work almost ready for release. This will include various things that are needed in a practical data management framework. (At least our IBuilder product needs them.)

- proof (find out what ground statements are behind an inference)

- forward chaining (find out what inferences depend on a statement) (backward chaining remains the main mode)

- repudiation (find out what ground statements are contradicted by a new statement)

- slightly smoother integration with typical RDF APIs (reduce the need to explicitly flag the part-of-speech of each resource with Node(), Class() and Property())

Second, I hope to create new generation of Xpetal that converts UML to OWL plus a series of related functions (instance validation, visualisation, "profiling") for use in the electric power industry. This would be a sponsored project and I will try and convince the sponsors to make it open source. I would like to use Jena and JGraph for this.

Regards,
Arnolf

Hi Arnold,

[apologies for the length!]

First of all: thanks for developing GraphPath! I have just begun to play with it, but I can see immediately that it is very well thought out and architected, and appears to benefit from your several years of experience in RDF and related fields (yes, I did some Googling of your work -- amazing how much you can find out! -- as part of my assessment of GraphPath ;).

I'm a Python developer, so most of the RDF/RDFS/OWL "aware" packages I've looked at were Python (e.g.: CWM [works mostly, but arg what ugly code! and not intended for production use anyway], Versa [nice, and I have great respect for the 4Suite group, but something about 4Suite puts me off ... perhaps because I think of XML as a necessary evil!], some others), and GraphPath to me seems the most elegant, most general, and most Pythonic. Kudos!

So I am now planning to use GraphPath in a framework that I am developing for use in implementing applications that need several features: "Intelligent PDM", knowledge-and-model-driven design, analysis, and systems engineering, and collaboration and tool/data integration. Yeah, I know -- quite ambitious! ;) I notice that you were (are?) participating in the OMG and attended some meetings Larry Johnson's MANTIS group, so I'm sure you are familiar with the general domain of such applications (engineering, CAD/CAE/CAM, etc.).

I'm always looking for potential collaborators for my project, which is open-source. I haven't decided whether to open a SourceForge project for it yet; I run my own CVS server and I'm happy to share the code with anyone who's interested. Anonymous checkout/update is available, and I use Twisted's (Kevin Turner's) CVSToys, which runs a "commits" mailing list that notifies subscribers when a commit happens, and I have another Mailman list for developer discussions (although it's just me at the moment because of recent funding cuts ... ;).

So let me know if you might be interested and I'll send you more explanation.

But that wasn't the immediate reason for this message. :) To help me in browsing the GraphPath code, I used epydoc to generate the API docs -- I'll send you a tar file, if you'd like (although it's pretty trivial to run epydoc yourself ;).

One thing I was thinking about that might be beneficial to both of us would be for me to add epydoc-style annotations (e.g., @param, @type, etc.) to the class and function docstrings so that the parameter lists and return values for everything would be documented and epydoc could generate a nicer and more complete API documentation. But you might probably have to answer some questions if I run into some that I can't figure out.

The other thing I'd offer to do (which is a matter of style, so it's your call, of course!! ;) would be to change all the tabs to 4 spaces and/or wrap all the lines to 80 characters. (Hope you're not offended at the boldness of this suggestion ... I'm a bit obsessive-compulsive about that stuff, but it's really just a matter of taste ... ;).

But as to the actual code, it looks beautiful to me.

I'll make one more mention of my app, because I see that you have also been involved in UML/XMI and conversions between that and RDF variants, which is also of great interest to me: a very important feature of the framework I'm developing will be import/export and mapping/conversion of data to/from (1) RDF/RDFS/OWL, (2) XMI (UML/MOF/etc), and (3) STEP (ISO 10303) data in various exchange formats (10303-21, XML, etc.). I know you are very familiar with the first 2 and *probably* somewhat familiar with STEP, or at least know of AP212 (10303-212), which is the most relevant to the electrical power domain.

The combination of GraphPath and rdflib will be great for (1), and probably all I need, I think. For (2), the best Python app I've seen is PEAK, and fortunately, that part of PEAK is usable without the rest of PEAK! ;) It is *very* well done (Phillip Eby is arguably the Python "king of meta" ;), being driven at the meta-level by the XMI model -- it can read in the XMI definition of any version of UML and create an importer based on that!

For (3), I'm currently planning to use "Express Engine", which is actually written in Lisp by a couple friends of mine, but I paid them to give it a minimal command-line interface so I'll call it as a process and tell it to read in a STEP schema and a Part 21 or STEP-XML file and give it to me in a simplified XML format (among other things). Express Engine is a SourceForge project: http://exp-engine.sourceforge.net/

There are a lot of things that I want to do as far as intelligently using/integrating/reasoning with data that comes from all 3 of the types of data sources mentioned above -- which very generally represent (1) enterprise knowledge and business rules, (2) software development and systems engineering, and (3) CAD/CAE/CAM/PDM/etc. I'm very curious to know if this is something you are interested in.

Well, enough for now -- sorry for bending your ear for so long, and I hope at least some of it was of interest!

Cheers,
Steve Waterbury
NASA/GSFC

Hi Brian,

You should be able to simply remove the import readline with no ill effects. I will make a new release that tests for a misssing readline module and uses the latest rdflib.

- Arnold


On Wed, 9 Jun 2004 05:42, QT - brian wrote:

--
Arnold deVos
adv@langdale.com.au
< replied-to message removed by QT >

I am having a few problems getting this to run and wondered if you could help.

I am running this on Windows XP which seems to favor rdflib over Redland which doesn't have windows binaries. However it appears the latest version of rdflib isn't compatible? I managed to track down the older versions and I think I am okay here however now I get the following error:
File "C:\Python23\Lib\site-packages\graphpath\util\testrunner.py", line 3, in ? import readline
ImportError: No module named readline

It would seem it is looking for the Unix specific readline module? Is GraphPath able to run under Windows?

Thanks for your help,

-brian

Hello Danny,

I hope you find the library fairly easy to use. If you have your own Python RDF API you should find GraphPath useful. (I have as many different RDF API's as I have projects.)

A note on rdflib and redland: I found rdflib to be very efficient (in terms of memory,cpu). Redland is no doubt good stuff but its python binding lets it down somewhat. I will look into updating to the new rdflib. I might also try adapting rdflib to redland in such a way as to get the best of both worlds.

BTW I saw your blog entry (with the snake!).

-Arnold

[John - the programming language is Python, the data language is RDF, see http://www.w3.org/TR/rdf-primer/ ]

Hi Arnold,

GraphPath looks really interesting, though I haven't yet been able to play with it - it looks like rdflib (2.0.3) has changed in a way that means it doesn't work any more:
No module named store.AbstractTripleStore

Anyhow I've got Redland downloading as I type, I'll try and get that all runnning today (might take a while as I usually do Python work on this Win2k machine rather than my Debian server).

I've been playing with an API myself - I wanted one for Smalltalk, but got distracted and have been building in Python instead, will port later. I'm new to Python so it's fairly slow going (I normally use Java). I'll have a go at writing an adapter, if for no better reason than to borrow your test cases ;-)

Hi guys,

Anyone cares to exlain me what is the this language?
I've got a task from school and I hardly know how to start it...

John