Skip to content

Instantly share code, notes, and snippets.

@pebbie
Last active August 30, 2022 02:53
Show Gist options
  • Star 4 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save pebbie/5aafdbb14921cadbad53011ef5f61165 to your computer and use it in GitHub Desktop.
Save pebbie/5aafdbb14921cadbad53011ef5f61165 to your computer and use it in GitHub Desktop.
visualize BGP triples in SPARQL query
import sys
import os.path as path
from rdflib import Namespace, XSD, RDF, RDFS, OWL
from rdflib.term import Variable, URIRef, BNode, Literal
from rdflib.plugins.sparql.parser import parseQuery
from rdflib.plugins.sparql.parserutils import prettify_parsetree
from rdflib.plugins.sparql import prepareQuery
from rdflib.paths import Path
import pprint
import pygraphviz as pgv
import argparse
BLANKNODES = []
def get_values(alg, vals):
for obj in alg:
k = list(obj.keys())[0]
v = obj[k]
lname = str(k)
if lname in vals:
vals[lname].append(v)
else:
vals[lname] = [v]
def find_triples(alg, vals):
# TODO: parse based on name attribute of the current parse tree node (alg)
# print(type(alg))
result = []
if isinstance(alg, list):
akg = alg
else:
akg = [alg]
for al in akg:
# pprint.pprint(al)
if hasattr(al, 'name'):
# print(al.name)
# pprint.pprint(al)
pass
ak = dict(al).keys()
for key in ak:
if key in ['PV', 'var', '_vars', 'datasetClause', 'expr', 'op', 'A', 'lazy', 'service_string']:
continue
elif key == 'res' and isinstance(al[key], list):
# values()
get_values(al[key], vals)
continue
elif key == 'value':
# pprint.pprint(al)
# print(al[key])
for var in al['var']:
vals[str(var)] = []
for value in al[key]:
if isinstance(value, list):
for var in al['var']:
tmpval = value.pop(0)
vals[str(var)].append(tmpval)
else:
vals[str(var)].append(value)
continue
elif key == 'term':
#print(type(al), al.name)
# print(al)
continue
elif key == 'triples':
# yield alg.triples
result += [al.triples]
continue
#print(f'opening {key}')
result += find_triples(al[key], vals)
return result
def get_prefix(NS, uri):
for k, v in sorted(NS.items(), key=lambda x: len(x[1]), reverse=True):
if uri.startswith(str(v)):
return k
return None
def get_local_name(NS, uri):
pref = get_prefix(NS, uri)
return pref+':'+str(uri).replace(NS[pref], '') if pref is not None else str(uri)
def get_label(NS, term):
tname = str(term)
if isinstance(term, Variable):
tname = '?' + tname
elif isinstance(term, URIRef):
tname = get_local_name(NS, term)
elif isinstance(term, BNode):
if tname not in BLANKNODES:
BLANKNODES.append(tname)
tname = '_:bn' + str(BLANKNODES.index(tname)+1)
elif isinstance(term, Path):
# print(term.n3())
if hasattr(term, 'arg'):
aname = get_local_name(NS, str(term.arg))
tname = tname.replace(str(term.arg), aname)
elif hasattr(term, 'args'):
for arg in term.args:
tname = tname.replace(str(arg), get_local_name(NS, arg))
elif hasattr(term, 'path'):
aname = get_local_name(NS, str(term.path))
tname = tname.replace(str(term.path), aname)
tname = tname[5:-1]
return tname
def set_node_attr(viz, term, termlabel):
n = viz.get_node(termlabel)
if isinstance(term, Variable):
n.attr['style'] = 'dashed'
elif isinstance(term, BNode):
n.attr['style'] = 'dotted'
elif isinstance(term, Literal):
n.attr['shape'] = 'box'
return n
if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument(
'-infile', help="text file containing SPARQL query (.rq)", required=True)
parser.add_argument(
'-outfile', help="image file containing the visualized BGP (.png)")
parser.add_argument(
'-dot', action='store_true',
help="when set, additionally create the dot file")
parser.add_argument(
'-ns', nargs='*',
help="namespace prefix definition in the format of <prefix>=<prefixURI> or a filename of a text file containing this format for each line",
default=[])
parser.add_argument(
'-verbose', action='store_true',
help="print out intemediate debug info")
parser.add_argument(
'-layout', help="layout prog to pass on graphviz",
default='dot',
choices=['dot', 'neato', 'circo', 'fdp', 'sfdp', 'twopi'])
parser.add_argument(
'-colorscheme', default='accent8',
help="default graphviz color scheme"
)
args = parser.parse_args()
if not path.exists(args.infile):
print(f"{args.infile} does not exists. exiting..")
sys.exit(1)
if args.outfile is None:
base, ext = path.splitext(args.infile)
args.outfile = base + '.png'
print(f"opening {args.infile}...")
with open(args.infile) as fq:
q = fq.read()
NS = {"rdf": str(RDF), "rdfs": str(
RDFS), "owl": str(OWL), "xsd": str(XSD)}
for nsdef in args.ns:
if path.exists(nsdef):
with open(nsdef) as fns:
for line in fns:
if line.startswith('#'):
# ignore comment
continue
prefix, nsURI = tuple(line.strip().split('='))
if prefix not in NS:
NS[prefix] = nsURI
else:
prefix, nsURI = tuple(nsdef.strip().split('='))
if prefix not in NS:
NS[prefix] = nsURI
pq = prepareQuery(
q, initNs=NS)
# pprint.pprint(pq.algebra)
# print(prettify_parsetree(pq.algebra))
for prefix, nsURI in [n for n in pq.prologue.namespace_manager.namespaces()]:
if prefix not in NS:
NS[prefix] = str(nsURI)
if args.verbose:
pprint.pprint(pq.algebra)
G = pgv.AGraph(directed=True)
G.node_attr.update(colorscheme=args.colorscheme)
G.edge_attr.update(colorscheme=args.colorscheme)
values = {}
tris = find_triples(pq.algebra, values)
# pprint.pprint(tris)
# exit()
if tris is not None:
for gid, trisgrp in enumerate(tris):
for s, p, o in trisgrp:
if args.verbose:
print(repr(s), repr(p), repr(o))
if not isinstance(p, URIRef):
print(type(p))
# get term labels
sname = get_label(NS, s)
pname = get_label(NS, p)
oname = get_label(NS, o)
# add triple
G.add_edge(sname, oname)
# customize edge attribute
edge = G.get_edge(sname, oname)
if 'color' not in dict(edge.attr).keys():
edge.attr['color'] = gid+1
edge.attr['label'] = pname
if isinstance(p, Variable):
edge.attr['style'] = 'dashed'
# customize node attribute
snode = set_node_attr(G, s, sname)
onode = set_node_attr(G, o, oname)
#print(sname, oname, gid+1)
if 'color' not in dict(snode.attr).keys():
snode.attr['color'] = gid+1
if 'color' not in dict(onode.attr):
onode.attr['color'] = gid+1
if len(values.keys()) > 0:
for var in values:
lname = str(var)
varname = '?' + lname
for value in values[lname]:
valname = get_label(NS, value)
G.add_edge(valname, varname)
edge = G.get_edge(valname, varname)
node = G.get_node(valname)
node.attr['shape'] = 'box'
edge.attr['style'] = 'dashed'
edge.attr['dir'] = 'none'
G.layout(prog=args.layout)
if args.dot:
base, ext = path.splitext(args.outfile)
dotfile = base + '.dot'
G.write(dotfile)
print(f"writing {dotfile}...")
print(f"writing {args.outfile}...")
G.draw(args.outfile)
@pebbie
Copy link
Author

pebbie commented Aug 24, 2020

next version, documentation & examples will be available in github repo https://github.com/pebbie/sparqlgpviz

@pebbie
Copy link
Author

pebbie commented Sep 12, 2020

Dear Peb Ruswono Aryan,

We have completed the review phase of ISWC 2020 Posters and Demos Track. On behalf of the ISWC 2020 Posters and Demos Track program committee, we regret to inform you that your submission

569: Peb Ruswono Aryan, Fajar J. Ekaputra, SPARQLGPViz: SPARQL Graph Pattern Visualization

could not be accepted for presentation.

This year, we received a high number of submissions for presentation at the conference. The competition was particularly tough due to the quality of submissions. For borderline papers, we have held discussions but sometimes had to make difficult decisions.

We know this is disappointing news and we understand that all authors have devoted a significant amount of effort to their work. So, we sincerely thank you for your effort.
We have included the reviews at the bottom of this email and we hope that the reviews can help you to further improve your work.

We look forward to seeing you at the virtual ISWC 2020 in November. You may register here:
https://iswc2020.semanticweb.org/attending/registration/. Information on student grants (not necessarily requiring paper acceptance) and a student mentoring session can be found at https://iswc2020.semanticweb.org/attending/student-grants/ and https://iswc2020.semanticweb.org/program/student-engagement-and-mentoring-session/

Thank you for submitting your work to ISWC.

Sincerely,
Kerry Taylor
Rafael Goncalves

SUBMISSION: 569
TITLE: SPARQLGPViz: SPARQL Graph Pattern Visualization

----------------------- REVIEW 1 ---------------------
SUBMISSION: 569
TITLE: SPARQLGPViz: SPARQL Graph Pattern Visualization
AUTHORS: Peb Ruswono Aryan and Fajar J. Ekaputra

----------- Reviewer's confidence -----------
SCORE: 4 ((high))
----------- Appropriateness -----------
SCORE: 2 (excellent)
----------- Outrageous -----------
SCORE: 0 (somewhat novel)
----------- Related work -----------
SCORE: 0 (fair)
----------- Novelty -----------
SCORE: 0 (fair)
----------- Uptake -----------
SCORE: 2 (poor)
----------- Impact -----------
SCORE: 0 (fair)
----------- Technical quality and depth -----------
SCORE: 0 (fair)
----------- Reusability -----------
SCORE: 1 (good)
----------- Reproducibility -----------
SCORE: 0 (fair)
----------- Design quality -----------
SCORE: 0 (fair)
----------- Justified Sem. Web use -----------
SCORE: 3 (fair)
----------- Availability and sustainability -----------
SCORE: 1 (good)
----------- Clarity and quality of writing -----------
SCORE: 0 (fair)
----------- Semantic technologies -----------
SCORE: 1 (good)
----------- Summary review -----------
The paper presents "SPARQLGPViz", a didactic tool to teach SPARQL queries constructs to non-technical users via visualization of their graph patterns. It describes the visual representation of basic SPARQL graph patterns that include URI nodes, variable type nodes, blank nodes, and literals. It's a preliminary work that could be useful in various use cases beyond SPARQL learning.
----------- Overall evaluation -----------
SCORE: 0 (borderline paper)
----- TEXT:
Positive remarks:

  • Nice examples presented at the Github project repository.

Negative remarks:

----------------------- REVIEW 2 ---------------------
SUBMISSION: 569
TITLE: SPARQLGPViz: SPARQL Graph Pattern Visualization
AUTHORS: Peb Ruswono Aryan and Fajar J. Ekaputra

----------- Reviewer's confidence -----------
SCORE: 4 ((high))
----------- Appropriateness -----------
SCORE: -2 (very poor)
----------- Outrageous -----------
SCORE: -2 (nothing new)
----------- Related work -----------
SCORE: -1 (poor)
----------- Novelty -----------
SCORE: -2 (very poor)
----------- Uptake -----------
SCORE: 1 (very poor)
----------- Impact -----------
SCORE: -2 (very poor)
----------- Technical quality and depth -----------
SCORE: -2 (very poor)
----------- Reusability -----------
SCORE: -1 (poor)
----------- Reproducibility -----------
SCORE: -1 (poor)
----------- Design quality -----------
SCORE: -2 (very poor)
----------- Justified Sem. Web use -----------
SCORE: 2 (poor)
----------- Availability and sustainability -----------
SCORE: -2 (very poor)
----------- Clarity and quality of writing -----------
SCORE: -2 (very poor)
----------- Semantic technologies -----------
SCORE: 0 (fair)
----------- Summary review -----------
The contribution is too low to be accepted.
----------- Overall evaluation -----------
SCORE: -3 (strong reject)
----- TEXT:

  • The authors present a collaborative multiview system to work on ontologies. There is an emphasis on adding multimodal content for Hyperknowledge relationships.
  • Paper: good structure, good writing, easy to follow
  • Theme: fits perfectly well to ISWC
  • Demo videos: three links to videos are given to present the web-based system. The system is shown based on a demo set for smart city contexts. The linkage of Hyperknowledge relationships is shown in video 2. A multiview system, for example shows the hierarchy and the actual graph.

This seems like an interesting demo for the ISWC audience.

----------------------- REVIEW 3 ---------------------
SUBMISSION: 569
TITLE: SPARQLGPViz: SPARQL Graph Pattern Visualization
AUTHORS: Peb Ruswono Aryan and Fajar J. Ekaputra

----------- Reviewer's confidence -----------
SCORE: 5 ((expert))
----------- Appropriateness -----------
SCORE: 2 (excellent)
----------- Outrageous -----------
SCORE: -2 (nothing new)
----------- Related work -----------
SCORE: 0 (fair)
----------- Novelty -----------
SCORE: -2 (very poor)
----------- Uptake -----------
SCORE: 3 (fair)
----------- Impact -----------
SCORE: 0 (fair)
----------- Technical quality and depth -----------
SCORE: 0 (fair)
----------- Reusability -----------
SCORE: 2 (excellent)
----------- Reproducibility -----------
SCORE: 2 (excellent)
----------- Design quality -----------
SCORE: -1 (poor)
----------- Justified Sem. Web use -----------
SCORE: 5 (excellent)
----------- Availability and sustainability -----------
SCORE: 1 (good)
----------- Clarity and quality of writing -----------
SCORE: 1 (good)
----------- Semantic technologies -----------
SCORE: 2 (excellent)
----------- Summary review -----------
The tool described is potentially useful and publicly sharing the code much appreciated,
but, in my opinion, this contribution is on itself of very limited research value, lacking any novelty.
----------- Overall evaluation -----------
SCORE: -2 (reject)
----- TEXT:
The authors describe a way of visualising graph patterns in SPARQL queries and a
tool that generates the visualisation for a given query.

This kind of visualisation is pretty typical and has been used in SPARQL visual
query builders. It is not clear if there is anything novel in the specific
visualisation proposed by the authors.
The tool is potentially useful and publicly sharing the code much appreciated,
but, in my opinion, this contribution is on itself of very limited research value.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment