This is a list of all current, planned, or imagined LarKC plugins. Please list your plugins, and describe each very briefly on its own page.
These plugins are used in a number of pipeline configurations which are described at the end of this page
Note:
For plugins under development, i.e. not yet working and not ready to be used/invoked by others, we suggest to set their status to: 'under development'
For plugins that are already developed and can be used/invoked by other plugins, i.e. they work, we suggest to set their status to: 'working'
Identify
plugin name |
responsible |
work package |
status |
description |
Sindice Triple Pattern Identifier (example) |
Barry Bishop and Mick Kerrigan |
working |
The Sindice Triple Pattern Identifier takes a set of Triple Patterns as inputs and returns a set of RDF Graphs as output that contain these triple patterns. This plugin performs its functionality by contacting the Sindice semantic web index. |
|
Sindice Keyword Identifier (example) |
Barry Bishop and Mick Kerrigan |
working |
The Sindice Keyword Identifiertakes a set of keywords as input and returns a set of RDF Graphs as output that contain these keywords. This plugin performs its functionality by contacting the Sindice semantic web index. |
|
Swoogle Document Search Identifier (example) |
Barry Bishop and Mick Kerrigan |
working |
The Swoogle Document Search Identifier takes a set of keywords as input and returns a set of RDF Graphs as output that contain these keywords. This plugin performs its functionality by contacting the Swoogle search engine and using its document search facility. |
|
Swoogle Ontology Search Identifier (example) |
Barry Bishop and Mick Kerrigan |
working |
The Swoogle Ontology Search Identifier takes a set of keywords as input and returns a set of RDF Graphs as output that contain these keywords. This plugin performs its functionality by contacting the Swoogle search engine and using its ontology search facility. |
|
Swoogle Term Search Identifier (example) |
Barry Bishop and Mick Kerrigan |
working |
The Swoogle Term Search Identifier takes a set of keywords as input and returns a set of RDF Graphs as output that contain these keywords. This plugin performs its functionality by contacting the Swoogle search engine and using its term search facility. |
|
?GeoLocationIdentifier |
CEFRIEL and CycEur |
working |
Interacting with an ?AllegroGraph Server this plugin selects triples located in a given area (defined with a center and a range). |
|
?RemoteGraphLoaderIdentifier |
CEFRIEL and CycEur |
working |
Loads a remote RDF file stored in a known location (its URL) |
|
Article Identifier |
Blaz Fortuna |
Beta |
Extracts list of URLs pointing to articles from the query and returns them as a collection of ?NaturalLanguageDocumentUrl |
Transform
plugin name |
responsible |
work package |
status |
description |
SPARQL to Triple Pattern Query Transformer (example) |
Barry Bishop and Mick Kerrigan |
working |
This transformer plugin takes a SPARQL query as input and returns a set of triple patterns that are contained within the SPARQL query. This transformer enables SPARQL queries to be used in conjunction with other plugins that take triple pattern queries as input. |
|
SPARQL to Keyword Query Transformer (example) |
Barry Bishop and Mick Kerrigan |
working |
This transformer plugin takes a SPARQL query as input and returns a list of keywords that are contained within the SPARQL query. This transformer enables SPARQL queries to be used in conjunction with other plugins that take keyword queries as input. |
|
Semantic Annotation Transformer |
USFD |
working |
This is a retrieval plugin that wraps an arbitrary semantic annotation pipeline from GATE. It has been written in response to the need for retrieving semantically annotations from text in several LarKC use case scenarios. NB currently some paths are hard-coded, awaiting changes to the LarKC platform API. |
|
RDF to Matrix Transformer |
Siemens |
under development |
Given a predefined key entity (e.g., student), this plugin transforms a set of RDF-triples related to the key entities to a 2-dimensional matrix in which the rows are the entities (students), the columns are the features of interest (sex, age, etc.). The entries are binary: a one stands for the existence of a corresponding RDF-triple in the triple set, and zero, otherwise. The generated matrix could be used as input to learning-based reasoner plugins. |
|
Inductive Materialization |
Siemens |
under development |
Estimation of triple probabilities using machine learning approaches. Storing of selected learned probabilistic triples in LarKC data layer for querying. |
|
Gate Transformer |
Blaz Fortuna |
Beta |
Takes an information set and extracts named-entities and stores them as a set of RDF statements |
|
XML2RDF Transformer |
Daniele Dell'Aglio (CEFRIEL) |
Beta |
Given an XML file and an XSL transformation, it extracts an RDF model from the XML document. |
Select
plugin name |
responsible |
work package |
status |
description |
?SingleGraphSelecter (example) |
Barry Bishop and Mick Kerrigan |
working |
This selecter takes a dataset as input and returns the first RDF graph within the dataset to be used for reasoning over. |
|
?PassItOnSelecter (example) |
Barry Bishop and Mick Kerrigan |
working |
This selecter will take a ?SetOfStatements and pass the very same ?SetOfStatements as output (in otherwords it provides a full selection over whats passed in) |
|
?GrowingDataSetSelecter (example) |
Barry Bishop and Mick Kerrigan |
working |
This selecter takes a ?SetOfStatements as input and incrementally builds a larger dataset each time the select method is called (in otherwords it provides a full selection over whats passed in going back over the history of all invocations of the select method) |
|
Baseline selector, corpus based (V2) |
USFD |
working |
Assumes that the selection context included a textual corpus which has been semantically annotated and has contributed a large number of triples to the knowledge store. This corpus is used to contribute lexical data upon which to apply standard information retrieval techniques |
|
Baseline selector (V3) |
Ontotext |
working |
Uses simple full-text search methods to extract parts of the dataset based on text tokens found in the query. It acts as a baseline naive implementation of selection. |
|
Key phrase selector |
Ontotext |
working |
Similar to the Baseline selector, it adapts selection criteria based only on characteristic key-phrases, instead of any text tokens appearing in the dataset. |
|
Spreading activation selector |
Onto |
working |
Select subsets of nodes using spreading activation based on weighted RDF graphs, weighting based on human models of cognition |
|
Prior Knowledge selector |
USFD |
working |
Given a set of ontology instance identifiers, a triple store repository, and some knowledge selection criteria, returns a subset of the repository that further describes the identifiers. This plugin is derived from a use case to prior knowledge about gene probes (SNPs) in a LarKC use case, but is generalised beyond this use case. |
|
Ranked selector |
USFD |
working |
Given a set of ontology instance identifiers and prior knowledge about the instances, rank these instances according to some model. In this case, we use Bayesian False Discovery Probability (BFDP), selecting and returning the top n ranked objects. Again, this is a general solution influenced by the specific LarKC use case of ranking gene probes (SNPs) based on their BFDP. |
|
RDF molecule based selection |
Ontotext |
planned |
A family of plugins in which selection is based on keywords in the RDF graph context of a node, using TFIDF, page rank, etc. |
|
Geometrical molecule selector |
Ontotext/USFD |
planned |
Geometrical model versions of the above |
|
Network statistics selector |
MPG |
planned |
Select a subset of nodes based on network measures of closeness and betweeness |
|
Interest-based selector |
WICI |
planned |
Select a subset of the knowledge source based on user interests. This plugin aims at removing scalability by diversity of user backgrounds. |
|
Cyc Selecter |
Blaz Fortuna and Luka Bradesko |
Beta |
Selects the companies with correspondence in ?ResearchCyc KB |
Reason
plugin name |
responsible |
WP |
status |
description |
M8 |
M11 |
M14 |
M17 |
M20 |
after M20 |
Deliverable |
SPARQL Query Evaluation Reasoner (example) |
Barry Bishop and Mick Kerrigan |
release |
This reasoner plugin wraps the Jena reasoner and enables the execution of SPARQL Select, Construct, Describe and Ask queries to be executed against it. |
|
|
|
|
|
|
|
|
?SparqlReasoner |
STI |
release |
a Jena reasoner with simple sparql query evaluation |
release |
|
official release with LarKC platform |
|
|
|
D4.7.1(M18) |
|
?PelletReasoner |
VUA |
release |
wrapper of Pellet SPARQL DL Reasoner |
|
release |
official release with LarKC platform |
|
|
|
D4.3.1 (M18) |
|
DIGReasoner |
VUA |
release |
DIG Reasoner interface which allows to call an external DIG reasoner, like RACER, FACT++,KOAN2, PION, etc. |
|
|
release |
|
|
|
D4.3.1 (M18) |
|
SimplePIONReasoner |
VUA |
release |
PION reasoner with syntactic relevance-based selection function without backtracking |
|
|
release with LarKC platform (?) |
|
|
|
D4.3.1 (M18) |
|
PIONReasoner |
VUA |
|
PION reasoner with syntactic relevance-based |
|
|
|
release |
|
|
D4.3.2(M24) |
|
PIONReasonerwithStopRules |
VUA |
|
PION reasoner by interleaving reasoning and selection with stop rules |
|
|
|
|
|
release |
D4.3.2(M24) |
|
CRIONReasoner |
VUA |
|
A reasoner supports contrastive reasoning with inconsistent ontologies |
|
|
|
|
|
alpha release |
D4.7.2(M33) |
|
?RuleReasoner |
STI |
|
A rule-based reasoner configurable with a variety of inference rules: RDF, RDFs, L2, etc. Later versions (after M14) will include approximate reasoning techniques |
|
|
prototype release |
|
|
|
D4.4.1(M24) design |
|
Cyc-based Decider |
?CycEuro |
working |
Wrapper of ?ResearchCyc |
|
release |
official release with LarKC platform |
|
|
|
D4.7.1(M18) |
|
Stream Reasoner |
CEFRIEL |
|
|
|
|
|
|
|
|
|
D4.6.1(M18) |
WebPIE Reasoner Plug-in |
VUA |
|
|
|
|
|
|
|
|
D4.7.2 (M33) |
|
?OpResPathFinderReasoner |
CEFRIEL and CycEur |
working |
A reasoner to apply graph algorithms (actually Dijkstra) to find the "desirable path" on graphs modeled with RDF (with a fixed schema) |
|
|
release |
|
|
|
D6.3 (M14) |
|
OBPathFinderReasoner |
Saltlux and Barry Bishop |
working |
A reasoner to compute the shortest path with ?OntoBroker on a graphs modeled in RDF |
|
|
release (?) |
|
|
|
|
|
Cyc Reasoner |
Blaz Fortuna and Luka Bradesko |
Beta |
Reasoner plug-in, which transfers reasoning requests to the ?ResearchCyc reasoner. Currently it have limited SPARQL to CycL / RDF to ?CycKb support |
||||||||
Interest-based Reasoner |
WICI |
Beta |
This plugin refine a SPARQL query by using user interests. The query results which are explicitly related with user interests will be provided. |
Decide
plugin name |
responsible |
work package |
status |
description |
Simple Configurable One Pass Decider (example) |
Barry Bishop and Mick Kerrigan |
working |
The Simple Configurable One Pass Decider plugin can be used to construct a simple linear decider from a set of prespecified larkc plugins. As the name suggests the decider will execute the entire pipeline through one iteration and return the result of the query as output. |
|
Simple Anytime Decider (example) |
Barry Bishop and Mick Kerrigan |
working |
The Simple Anytime Decider is a simple example plugin from the baby larkc demo. This decider executes a pipeline iteratively over a set of example plugins, returning results at the end of each iteration. Thus this decider returns intermediate results during its execution before returning its final result once the pipeline is exhausted. |
|
UrbanBabyLarKC Decider |
CEFRIEL and CycEur |
working |
The UrbanBabyLarKC Decider allows to execute a pipeline to compute paths between two fixed nodes (that represents a start and a goal points in Milan). The pipeline is computed iteratively in order to try find the best path |
|
Cyc-based Decider |
Luka Bradesko, Blaz Fortuna |
under development |
Uses metareasoning and Qos parameters of the plug-ins to construct the pipelines |
Pipeline configurations that use these plugins
"SPARQL for the Web" pipeline:This pipeline takes a SPARQL query, uses Sindice to retrieve relevant datasets for this query, uploads those datasets, and answers the query, all of this in a streaming fashion. Plugins are:
TRANSFORM: SPARQL to Triple Pattern Query Transformer or SPARQL to Keyword Query Transformer
IDENTIFY: Sindice Triple Pattern Identifier or Sindice Keyword Identifier
SELECT: Growing ?DataSet Selecter
REASON: SPARQL Query Evaluation Reasoner
DECIDE: Simple Configurable One Pass Decider or Cyc-based Decider
"NewsSPARQL" ("Cyc-Gate") pipeline:This pipeline takes a news article (from disk, or from Google News), runs GATE for named entity recognition, maps these named entities to items in the Cyc knowledgebase, and then uses Cyc to answer questions. For example "which software companies where in the news yesterday, and what are their stockmarket names".
IDENTIFY: Article Identifier
TRANSFORM: Gate Transformer
SELECT: Cyc Selecter
REASON: Cyc Reasoner
"alpha Urban LarKC" pipeline: this is the pipeline of the Urban Computing demo at the first review. It looks for relevant monuments in Milano and events taking place today in Milano, and it finds a path to reach one of thos destinations starting from the current location of the user. The pipeline is constituted by a main decider, which, on the basis of the input query, runs a specific sub-pipeline to get the results, as follows:
DECIDE: alpha Urban LarKC decider
subpipeline 1: monument destination selection
IDENTIFY: Urban Sindice Identifier, which finds relevant RDF documents containing info about Milano documents, by querying Sindice
SELECT: Growing ?DataSet Selecter
REASON: Sparql Query Evaluation Reasoner
subpipeline 2: event destination selection
IDENTIFY: Event Identifier, which invokes Eventful REST API to get info about today events in Milano
TRANSFORM: XML2RDF Transformer , which translates the XML results of the previous plugin into an RDF form by applying an XSL tranformation
SELECT: Growing ?DataSet Selecter
REASON: Sparql Query Evaluation Reasoner
subpipeline 3: path finding