Lightweight Semantic Annotation of Geospatial RESTful Services
1 like531 views
This document proposes a lightweight approach to semantically annotating RESTful services. It describes annotating the parameters of 60 sample RESTful services by matching them to DBpedia and Geonames ontologies or using suggestion services. It reports results of annotating 369 initial parameters with 191 matches, increasing to 1573 parameters and 159 matches using additional techniques. The approach was able to annotate 78 input parameters and 315 output parameters of the services and validate 56 of 60 services based on executions.
Lightweight Semantic Annotation of Geospatial RESTful Services
- 1. Lightweight Semantic Annotation of Geospatial RESTful Services Victor Saquicela, Luis. M. Vilches-Blazquez, and Oscar Corcho Facultad de Informática, Universidad Politécnica de Madrid Campus de Montegancedo s/n, 28660 Boadilla del Monte, Madrid http://www.oeg-upm.net {vsaquicela,lmvilches,ocorcho}@fi.upm.es Phone: 34.91.3366605, Fax: 34.91.3524819 Presented by Boris Villazón-Terrazas ESWC 2011 Project: EV Date: 02/06/2011 Heraklion, Greece
- 2. ToC • Introduction & motivation • Related work • Semantic annotation of RESTful services • A sample set of RESTful services • Syntactic description y p • Semantic annotation • Checking the semantic annotation of RESTful services Checking the semantic annotation of RESTful • Experimental results • Conclusions and future work 2
- 3. Introduction and motivation RESTful RESTful Service Y RESTful Service X Service Z RESTful Service S i W … Description in HTML that contains: • a list of the available operations RESTful • URIs and parameters Service W • expected output • error messages • set of examples of execution • 3274 of RESTful services (APIs) in programmableweb.com
- 4. Introduction • Description in HTML page 4
- 5. Related work • Syntactic • H Heavyweight: WSDL i ht • Lightweight: WADL, hREST • Semantic • Heavyweight: OWL-S, WSMO y g • Lightweight: SA-REST, MicroWSMO • Tools • SWEET • WSMO editor 5
- 6. Our approach Semantic annotation of RESTf l services b S ti t ti f RESTful i by (1) Obtaining and formalising their syntactic descriptions (2) Interpreting, and semantically enriching their parameters Our approach uses diverse types of resources • cross-domain ontology, DBpedia (combined with GeoNames) • Suggestion services • Synonym services Semantic annotation Repository Syntactic description 6
- 7. A sample set of RESTful services • Service 1. retrieves information related to a country. • http://ws.geonames.org/countryInfo?country=ES • Service 2. retrieves information about places (venues). • htt // i http://api.eventful.com/rest/venues/search?app_key=p4t8BFcLDt C tf l / t/ / h? k 4t8BF LDt CzpxdS&location=Madrid dS&l ti M d id Service 1 Service 2 <venue id="V0-001-000154997-6"> <geonames> <url>http://eventful.com/madrid/venues/la- <country> y ancha-/V0-001-000154997-6</url> <countryCode>ES</countryCode> <country_name>Spain</country_name> <countryName>Spain</countryName> <name>La Ancha</name> <venue_name>La Ancha</venue_name> <isoNumeric>724</isoNumeric> <description></description> <isoAlpha3>ESP</isoAlpha3> <venue_type>Restaurant</venue_type> <fipsCode>SP</fipsCode> <address></address> <continent>EU</continent> <city_name>Madrid</city_name> <capital>Madrid</capital> <region_name></region_name> <region_abbr></region_abbr> <areaInSqKm>504782.0</areaInSqKm> <postal_code></postal_code> <population>40491000</population> <country_abbr2>ES</country_abbr2> <currencyCode>EUR</currencyCode> C d EUR / C d <country_abbr>ESP</country_abbr> <languages>es-ES,ca,gl,eu</languages> <longitude>-3.68333</longitude> <geonameId>2510769</geonameId> <latitude>40.4</latitude> <geocode_type>City Based GeoCodes <bBoxWest>-18.169641494751</bBoxWest </geocode_type> <bBoxNorth>43.791725</bBoxNorth> <owner>frankg</owner> owner frankg /owner <bBoxEast>4.3153896</bBoxEast> <timezone></timezone> <bBoxSouth>27.6388</bBoxSouth> <created></created> </country> <event_count>0</event_count> <trackback_count>0</trackback_count> </geonames> <comment_count>0</comment_count> <link_count>0</link_count> <li k t>0</li k t> <image></image> </venue> <venue id="V0-001-000154998-5"> 7
- 8. Syntactic description: Invocation and registration • User adds the URL of a service in our system • System invokes the RESTful service with some sample y parameters, by using SDO API • System analyzes the response (XML) to obtain a syntactic description of the parameters • Results • Service1. countryInfo($country,bBoxSouth,isoNumeric,continent,fipsCode,areaInSqK m,languages,isoAlpha3,countryCode,bBoxNorth,population,bBoxWest,curre ncyCode,bBoxEast,capital,geonameId,countryName) ncyCode bBoxEast capital geonameId countryName) • Service2. rest/venues/search($location,$app_key,id,link_count,page_count,longitude,t rackback_count,version,venue_type,owner,url,country_name,event_count,t rackback count,version,venue type,owner,url,country name,event count,t otal_items,city_name,address,name,latitude,page_number,postal_code,cou ntry_abbr,first_item,page_items,last_item,page_size,country_abbr2,comme nt_count,geocode_type,search_time,venue_name) • Results are stored and WADL file is generated. 8
- 9. Semantic annotation (I) • The starting point is the list of syntactic parameters • Th system retrieves all th classes/properties f The t ti ll the l / ti from DBpedia ontology whose names have a match with each p parameter of the RESTFul Service. • Exact match • Similarity measures: Jaro, Jaro Winkler, and Levenshtein • If the system obtains correspondences from the matching process, process it uses DBPedia concepts/properties to retrieve samples (RDF instances). • If there is no “matching”, we use external services • Spelling suggestion services • Synonym services 9
- 10. Semantic annotation: example Service1. countryInfo($country,bBoxSouth,isoNumeric,continent,fipsCode,areaInSqKm,languages, isoAlpha3,countryCode,bBoxNorth,population,bBoxWest,currencyCode,bBoxEast,capital, geonameId,countryName) geonameId countryName) Enriching the original parameters with external resources 10
- 11. Semantic annotation: more examples 11
- 12. Result Service1 (i1, i2, o1, o2, o3, o4, o5, o6, o7) Service 1 (http://dbpedia.org/ontology/Country,http,http://www.w3.org/2003/01/geo/wgs84_pos#lat,http://ww w.w3.org/2003/01/geo/wgs84_pos#long,isoNumeric,http://dbpedia.org/ontology/Continent,fipsCod e,http://dbpedia.org/property/areaMetroKm,languages,isoAlpha3,http://dbpedia.org/ontology/countr y,http://www.w3.org/2003/01/geo/wgs84_pos#lat,http://www.w3.org/2003/01/geo/wgs84_pos#long, http://dbpedia.org/ontology/populationDensity,http://www.w3.org/2003/01/geo/wgs84_pos#lat,http:// www.w3.org/2003/01/geo/wgs84_pos#long,http://dbpedia.org/ontology/Currency,http://www.w3.org /2003/01/geo/wgs84_pos#lat,http://www.w3.org/2003/01/geo/wgs84_pos#long,http://dbpedia.org/on tology/capital,geonameId,http://dbpedia.org/ontology/country) 12
- 13. Checking the semantic annotation (Input) • The system selects for each parameter a random subset of the example instances. • It makes several invocations of the service, by iterating over the registered values. g • If the service return results from the invocation -> the service is considered as executable; non-executable if it cannot invoked succesfully. Country http://dbpedia.org/ontology/country http://dbpedia.org/ontology/locationCountry Instance1 Instance1 Instance2 Instance2 Instance3 Instance3 ......... ......... InstanceM InstanceM 13
- 14. Checking the semantic annotation (Output) • The system only takes into account executions with correct inputs. • The system compares the outputs obtained after execution with the information already stored in the y repository. • If the output can be matched, the system considers the output annotations as valid. • Finally, the correspondences between the different parameters of the service and DBpedia/Geonames ontology are registered and stored into the respository. 14
- 15. Experimental results (I) • 60 RESTful services selected from www.programmableweb.com (manually) • 56 registered in the system • List of 369 different parameters, 52 input parameters and 342 output parameters, without duplications. Additional Matches (DBpedia and Attributes Total parameters GeoNames ontologies) Initial parameters 369 - 191 Parameters + Suggestions 426 57 33 Parameters + Synonyms y y 1573 1147 126 Parameters + Suggestions + Synonyms 1573 1204 159 15
- 16. Experimental results (II) Annotated Annotated RESTful Total Annotated Special Service parameters parameters Service parameters parameters parameters validation ( p (DBpedia)) ( (GeoNames) ) Input 152 78 33 45 73 56 4 parameters Output 862 315 202 113 299 - parameters Results of the input and output parameters Found Not found Not Right RESTful Service Annotated Precision Recall parameters parameters annotated parameters Output 475 92 315 160 242 0.66 0.77 parameters Output parameters metrics 16
- 17. Conclusions and future work • We have proposed an approach to perform semantic annotation process of RESTful services, using ontologies (DBpedia and Geonames) and external resources (suggestions and synonyms). • Future work: • GUI that will ease the introduction of existing service by users users. • Improving the SPARQL queries • More specific domain ontologies in the semantic process • Standards generation such as hREST, SA-REST and MicroWSMO • Integration with iServe. 17
- 19. Lightweight Semantic Annotation of Geospatial RESTful Services Victor Saquicela, Luis. M. Vilches-Blazquez, and Oscar Corcho Facultad de Informática, Universidad Politécnica de Madrid Campus de Montegancedo sn, 28660 Boadilla del Monte, Madrid http://www.oeg-upm.net {vsaquicela,lmvilches,ocorcho}@fi.upm.es Phone: 34.91.3366605, Fax: 34.91.3524819 Presented by Boris Villazón-Terrazas ESWC 2011 Project: EV Date: 02/06/2011 Heraklion, Greece