Wouldn’t it be nice to have an interactive environment where you quickly hack together an RDF model and then show it to your clients or colleagues in a more accessible format – i.e. a diagram?

Don’t know if there’s anything like that already, but the other day while polishing up the OntosPy library I’ve taken a couple of hours of fun-coding and put together a module that lets you do that.

The idea is simple: load an interactive environment where you can quickly sketch out a few ideas using the (very readable) Turtle rdf format.

Then export it onto a different representation e.g. a graphical one, so that it can be shown to people. Or just to keep working on it via a medium that offers different affordances.

So here it is:

[michele.pasin]@here:~/code/python>sketch.py 
Good morning. Ready to Turtle away. Type docs() for help.
In [1]: docs()

====Sketch v 0.2====

add()  ==> add statements to the graph
...........SHORTCUTS:
...........'class' = owl:Class
...........'sub' = rdfs:subClassOf
...........TURTLE SYNTAX:  http://www.w3.org/TR/turtle/

show() ==> shows the graph. Can take an OPTIONAL argument for the format.
...........eg one of['xml', 'n3', 'turtle', 'nt', 'pretty-xml', dot']

clear()	 ==> clears the graph
...........all triples are removed

omnigraffle() ==> creates a dot file and opens it with omnigraffle
...........First you must set Omingraffle as your system default app for dot files!

quit() ==> exit

====Have fun!====


In [2]: add()
Multi-line input. Enter ### when finished.
:person a class
:mike a :person
:person sub :agent
:organization sub :agent
:worksIn rdfs:domain :person
:worksIn rdfs:range :organization
:mike :worksIn :DamageInc
:DamageInc a :organization

In [3]: show()
@prefix : <http://this.sketch#> .
@prefix bibo: <http://purl.org/ontology/bibo/> .
@prefix foaf: <http://xmlns.com/foaf/0.1/> .
@prefix npg: <http://ns.nature.com/terms/> .
@prefix npgg: <http://ns.nature.com/graphs/> .
@prefix npgx: <http://ns.nature.com/extensions/> .
@prefix owl: <http://www.w3.org/2002/07/owl#> .
@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
@prefix skos: <http://www.w3.org/2004/02/skos/core#> .
@prefix xml: <http://www.w3.org/XML/1998/namespace> .
@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .

:mike a :person ;
    :worksIn :DamageInc .

:worksIn rdfs:domain :person ;
    rdfs:range :organization .

:DamageInc a :organization .

:organization rdfs:subClassOf :agent .

:person a owl:Class ;
    rdfs:subClassOf :agent .



In [4]: show("xml")
<?xml version="1.0" encoding="UTF-8"?>
<rdf:RDF
   xmlns="http://this.sketch#"
   xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
   xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#"
>
  <rdf:Description rdf:about="http://this.sketch#mike">
    <rdf:type rdf:resource="http://this.sketch#person"/>
    <worksIn rdf:resource="http://this.sketch#DamageInc"/>
  </rdf:Description>
  <rdf:Description rdf:about="http://this.sketch#organization">
    <rdfs:subClassOf rdf:resource="http://this.sketch#agent"/>
  </rdf:Description>
  <rdf:Description rdf:about="http://this.sketch#DamageInc">
    <rdf:type rdf:resource="http://this.sketch#organization"/>
  </rdf:Description>
  <rdf:Description rdf:about="http://this.sketch#person">
    <rdf:type rdf:resource="http://www.w3.org/2002/07/owl#Class"/>
    <rdfs:subClassOf rdf:resource="http://this.sketch#agent"/>
  </rdf:Description>
  <rdf:Description rdf:about="http://this.sketch#worksIn">
    <rdfs:domain rdf:resource="http://this.sketch#person"/>
    <rdfs:range rdf:resource="http://this.sketch#organization"/>
  </rdf:Description>
</rdf:RDF>

In [5]: omnigraffle()
### saves a dot file and tries to open it with your default editor
### if you're on a mac and have omnigraffle - that could be the one!

In [6]: quit()

If you are mac based and you have associated .dot files to the excellent Omnigraffle app, you’d see something like this:

That speeded up my work quite a bit – especially in situations where you don’t mind about precision but are more interested in quickly showing the merits of a modelling approach.

Any comments or ideas on how to develop this further?