The London Perl and Raku Workshop takes place on 26th Oct 2024. If your company depends on Perl, please consider sponsoring and/or attending.

SYNOPSIS

A Perl interpretation of the SimpleGraph developed in Python by Toby Segaran in his book "Programming the Semantic Web", published by O'Reilly, 2009. CPAN modules are used in place of the Python standard library modules used by Mr. Segaran.

my $graph = VANAMBURG::SEMPROG::SimpleGraph->new();

$graph->load("data/place_triples.txt");

$graph->add("Morgan Stanley", "headquarters", "New_York_New_York");

my @sanfran_key = $graph->value(undef,'name','San Francisco');

my @sanfran_triples = $graph->triples($sanfram_key, undef, undef);

my @bindings = $g->query([
   ['?company', 'headquarters', 'New_York_New_York'],
   ['?company', 'industry',     'Investment Banking'],
   ['?contrib', 'contributor',  '?company'],
   ['?contrib', 'recipient',    'Orrin Hatch'],
   ['?contrib', 'amount',       '?dollars'],
]);

for my $binding (@bindings){
   printf "company=%s, contrib=%s, dollars=%s\n", 
       ($binding->{company},$binding->{contrib},$binding->{dollars});
}


$graph->applyinference( VANAMBURG::SEMPROG::GeocodeRule->new() );

SimpleGraph

This module and it's test suite is inspired by the simple triple store implimentation developed in chapters 2 and 3 of "Programming the Semantic Web" by Toby Segaran, Evans Colin, Taylor Jamie, 2009, O'Reilly. Mr. Segaran uses Python and it's standard library to show the workins of a triple store. This module and it's test make the same demonstration using Perl and CPAN modules, which may be thought of as a Perl companion to the book for readers who are interested in Perl. Copies of Mr. Segaran's test data files are included in this distribution for your convenience.

In addition to SimpleGraph, the triple store, the other exercises presented in chapters 2 and 3 are here interpreted as a set of perl test programs, using Test::More and are found in the modules 't/' directory.

Triple Store Modules

lib/VANAMBURG/SEMPROG/SimpleGraph.pm

lib/VANAMBURG/SEMPROG/CloseToRule.pm
lib/VANAMBURG/SEMPROG/GeocodeRule.pm
lib/VANAMBURG/SEMPROG/InferenceRule.pm    
lib/VANAMBURG/SEMPROG/TouristyRule.pm
lib/VANAMBURG/SEMPROG/WestCoastRule.pm

Module Usage Shown in Tests

t/semprog_ch02_03_places.t
t/semprog_ch02_04_celebs.t
t/semprog_ch02_05_business.t
t/semprog_ch02_moviegraph.t
t/semprog_ch03_01_queries.t
t/semprog_ch03_02_inference.t
t/semprog_ch03_03_chain_of_rules.t
t/semprog_ch03_04_shortest_path.t
t/semprog_ch03_05_join_graph.t
qt/semprog_ch03_chain_of_rules.t

Find out more about, or get the book at http://semprog.com, the Semantic Programming web site.

INSTALLATION NOTES

This module can be installed via cpan. This method resolves dependency issues and is convenient. In brief, it looks something like this in a terminal on linux:

$sudo cpan
cpan>install VANAMBURG::SEMPROG::SimpleGraph
...
cpan>quit
$

All dependencies, as well as the modules are now installed. Leave out 'sudo' if using Strawberry perl on Windows.

You can then download the source package and read and run the test programs.

$tar xzvf VANAMBURG-SEMPROG-SimpleGraph-0.001.tar.gz
$cd VANAMBURG-SEMPROG-SimpleGraph-0.001/  
$ perl Makefile.PL
...
$make
...

Run 'dmake' instead of 'make' if using Strawberry Perl on Windows.

To run all the test programs:

$make test

-- Note that some tests require internet access for geo code data.

To run one test:

$prove -Tvl lib - t/semprog_ch03_05_join_graph.t 

MooseX::Declare Experiment

Version 0.007 was an experiment in using MooseX::Declare. The code remaind the same as version 0.006, except that classes were defined by the 'class' keyword instead of 'package' and methods are defined using 'method' keyword and well defined parameter lists in place of 'sub' and '@_'. 'class' and 'method' are supplied by MooseX::Declare.

Types of Changes to Source Files

The types of changes to the source looks like this.

1) CLASS DECLARATIONS WERE CHANGED

OLD PACKAGE STATEMENTS REMOVED:
<<package VANAMBURG::SEMPROG::SimpleGraph;
<<use Moose;

REPLACED WITH MUCH CLEANER DECLARATIONS:
>>use MooseX::Declare;
>>class VANAMBURG::SEMPROG::SimpleGraph{

2) METHOD DECLARATIONS WERE CHANGED

OLD SUB AND @ARG REMOVED:
<<sub _addToIndex{
<<    my ($self, $index, $a, $b, $c) = @ARG;
 
REPLACED WITH METHOD AND DEFINED OPTIONAL PARAMS:
>>method add($sub?, $pred?, $obj?){

Performance Changes

Version 0.007, using MooseX::Declare took ten times as long as using Moose alone.Subsequent to this test, version 0.008 was created by rolling back to the Version 0.006 sources.

Devel::NYTProf For Version 0.006 (Moose only)

Performance Profile Index
For t/semprog_ch03_01_queries.t
  Run on Sun Jan 10 01:14:07 2010
Reported on Sun Jan 10 01:16:44 2010

Profile of t/semprog_ch03_01_queries.t for 64.0s, executing 10616075 statements and 3106325 subroutine calls in 119 source files and 194 string evals.
Top 15 Subroutines — ordered by exclusive time 
Calls 	P 	F 	Exclusive
Time 	Inclusive
Time 	Subroutine
374201	2	2	6.10s	8.74s	Set::Scalar::Base::::_insert_elements Set::Scalar::Base::_insert_elements
187100	1	1	3.90s	15.2s	Set::Scalar::::_insert_hook Set::Scalar::_insert_hook
451618	2	2	3.18s	3.18s	Set::Scalar::Base::::_invalidate_cached Set::Scalar::Base::_invalidate_cached
264518	4	3	2.61s	3.88s	Set::Scalar::Base::::_make_elements Set::Scalar::Base::_make_elements
109683	3	1	2.45s	26.6s	VANAMBURG::SEMPROG::SimpleGraph::::_addToIndexVANAMBURG::SEMPROG::SimpleGraph::_addToIndex
109683	2	1	1.86s	14.0s	Set::Scalar::Real::::insert Set::Scalar::Real::insert
187100	2	2	1.85s	17.1s	Set::Scalar::Base::::_insert Set::Scalar::Base::_insert
187101	2	2	1.72s	6.10s	Set::Scalar::Virtual::::_extend Set::Scalar::Virtual::_extend
77417	1	1	1.63s	9.01s	Set::Scalar::::_new_hook Set::Scalar::_new_hook
36561	1	1	1.39s	28.6s	VANAMBURG::SEMPROG::SimpleGraph::::addVANAMBURG::SEMPROG::SimpleGraph::add
77417	1	1	1.36s	1.90s	Set::Scalar::Real::::_delete Set::Scalar::Real::_delete
77417	1	1	1.32s	6.82s	Set::Scalar::Real::::clear Set::Scalar::Real::clear
219366	1	1	1.26s	1.26s	Set::Scalar::Base::::_strval Set::Scalar::Base::_strval
77417	1	1	1.11s	4.73s	Set::Scalar::Real::::delete Set::Scalar::Real::delete
77419	3	2	1.11s	10.1s	Set::Scalar::Base::::new Set::Scalar::Base::new

Devel::NYTProf For Version 0.007 (MooseX::Declare)

Performance Profile Index
For t/semprog_ch03_01_queries.t
  Run on Sun Jan 10 01:28:09 2010
Reported on Sun Jan 10 01:38:25 2010

Profile of t/semprog_ch03_01_queries.t for 489s, executing 74793743 statements and 24836936 subroutine calls in 371 source files and 407 string evals.
Top 15 Subroutines — ordered by exclusive time 
Calls 	P 	F 	Exclusive
Time 	Inclusive
Time 	Subroutine
1426306	23	11	38.3s	250s	MooseX::Types::TypeDecorator::::AUTOLOAD MooseX::Types::TypeDecorator::AUTOLOAD
438753	1	1	36.5s	186s	MooseX::Types::Structured::::__ANON__[MooseX/Types/Structured.pm:745] MooseX::Types::Structured::__ANON__[MooseX/Types/Structured.pm:745]
2304288	3	1	26.0s	30.6s	MooseX::Types::TypeDecorator::::__type_constraint MooseX::Types::TypeDecorator::__type_constraint
1280098	6	5	21.2s	163s	Moose::Meta::TypeConstraint::::check Moose::Meta::TypeConstraint::check
548425	2	2	15.1s	18.8s	Moose::Meta::TypeConstraint::::Defined Moose::Meta::TypeConstraint::Defined
146251	1	1	12.2s	16.8s	MooseX::Method::Signatures::Meta::Method::::__ANON__[MooseX/Method/Signatures/Meta/Method.pm:430] MooseX::Method::Signatures::Meta::Method::__ANON__[MooseX/Method/Signatures/Meta/Method.pm:430]
585005	2	1	11.2s	303s	MooseX::Meta::TypeConstraint::Structured::::__ANON__[MooseX/Meta/TypeConstraint/Structured.pm:115] MooseX::Meta::TypeConstraint::Structured::__ANON__[MooseX/Meta/TypeConstraint/Structured.pm:115]
1426306	1	2	9.06s	9.06s	MooseX::Types::TypeDecorator::::CORE:match MooseX::Types::TypeDecorator::CORE:match (opcode)
438881	9	7	8.41s	17.2s	MooseX::Types::TypeDecorator::::isa MooseX::Types::TypeDecorator::isa
3924871	56	34	7.80s	7.80s	Scalar::Util::::blessed Scalar::Util::blessed (xsub)
1426685	7	6	7.22s	7.22s	Moose::Meta::TypeConstraint::::_compiled_type_constraint Moose::Meta::TypeConstraint::_compiled_type_constraint
374201	2	2	6.31s	9.05s	Set::Scalar::Base::::_insert_elements Set::Scalar::Base::_insert_elements
219366	4	2	5.35s	227s	VANAMBURG::SEMPROG::SimpleGraph::::_addToIndex VANAMBURG::SEMPROG::SimpleGraph::_addToIndex
146251	1	1	4.87s	256s	MooseX::Meta::TypeConstraint::ForceCoercion::::validateMooseX::Meta::TypeConstraint::ForceCoercion::validate
146251	1	1	4.87s	263s	MooseX::Method::Signatures::Meta::Method::::validate MooseX::Method::Signatures::Meta::Method::validate

METHODS

add

Adds a triple to the graph.

$g->add("San Francisco", "inside", "California");
$g->add("Ann Arbor", "inside", "Michigan");

remove

Remove a triple pattern from the graph.

# remove all triples with predicate "inside"
$g->remove(undef, "inside", undef);

triples

# retrieve all triples with predicate "inside"
my @triples = $g->triples(undef, "inside", undef);

# @triples looks like this:
#  ( 
#    ["San Francisco", "inside", "California"],
#    ["Ann Arbor", "inside", "Michigan"],
#  )

value

Retrieve a single value from a triple.

my $x = $g->value(undef, 'inside', 'Michigan');
# $x contains "Ann Arbor" given examples added.

query

Returns array of hashrefs where keys are binding variables for triples.

    my @bindings = $g->query([
	['?company','headquarters','New_York_New_York'],
	['?company','industry','Investment Banking'],
	['?cont','contributor','?company'],
	['?cont', 'recipient', 'Orrin Hatch'],
	['?cont', 'amount', '?dollars'],
    ]);

applyinference

Given an InferenceRule, generates additional triples in the triple store.

load

Loads a csv file in utf8 encoding.

$g->load("some/file.csv");

load_json

Loads a json file into a graph. The json file should be formated as follows:

{ "triples" : [ { "s": "your subject 1", "p": "your predicate 1", "o": "your object 1" }, { "s": "your subject 2", "p": "your predicate 2", "o": "your object 2" } ] }

save

Saves a csv file in utf8 encoding.

$g->load("some/file.csv");

_addToIndex

See source for details.

_removeFromIndex

Removes a triple from an index and clears up empty indermediate structures.

1 POD Error

The following errors were encountered while parsing the POD:

Around line 469:

Non-ASCII character seen before =encoding in '—'. Assuming UTF-8