use of org.apache.stanbol.entityhub.servicesapi.query.FieldQuery in project stanbol by apache.
the class EntitySearcherUtils method createFieldQuery.
/**
* Validated the parsed parameter as parsed to
* {@link EntitySearcher#lookup(String, Set, List, String...)}
* and creates a fieldQuery for the parsed parameter
* @param field
* @param includeFields
* @param search
* @param languages
* @return
*/
public static final FieldQuery createFieldQuery(FieldQueryFactory factory, String field, Set<String> includeFields, List<String> search, String... languages) {
if (field == null || field.isEmpty()) {
throw new IllegalArgumentException("The parsed search field MUST NOT be NULL nor empty");
}
if (search == null || search.isEmpty()) {
throw new IllegalArgumentException("The parsed list of search strings MUST NOT be NULL nor empty");
}
// build the query and than return the result
FieldQuery query = factory.createFieldQuery();
if (includeFields == null) {
query.addSelectedField(field);
} else {
if (!includeFields.contains(field)) {
query.addSelectedField(field);
}
for (String select : includeFields) {
query.addSelectedField(select);
}
}
// TODO make configurable
query.setLimit(20);
query.setConstraint(field, new TextConstraint(search, languages));
return query;
}
use of org.apache.stanbol.entityhub.servicesapi.query.FieldQuery in project stanbol by apache.
the class ReferencedSiteSearcher method lookup.
@Override
public Collection<? extends Representation> lookup(String field, Set<String> includeFields, List<String> search, String... languages) throws IllegalStateException {
// build the query and than return the result
Site site = getSearchService();
if (site == null) {
throw new IllegalStateException("ReferencedSite " + siteId + " is currently not available");
}
FieldQuery query = EntitySearcherUtils.createFieldQuery(site.getQueryFactory(), field, includeFields, search, languages);
if (limit != null) {
query.setLimit(limit);
}
QueryResultList<Representation> results;
try {
results = site.find(query);
} catch (SiteException e) {
throw new IllegalStateException("Exception while searchign for " + search + '@' + Arrays.toString(languages) + "in the ReferencedSite " + site.getId(), e);
}
return results.results();
}
use of org.apache.stanbol.entityhub.servicesapi.query.FieldQuery in project stanbol by apache.
the class BaseGoogleRefineReconcileResource method reconcile.
private JSONObject reconcile(String siteId, ReconcileQuery rQuery) throws JSONException, EntityhubException {
FieldQuery query = createFieldQuery(siteId);
query.addSelectedFields(SELECTED_FIELDS);
addNameConstraint(rQuery, query);
addTypeConstraint(rQuery, query);
addPropertyConstraints(rQuery, query);
query.setLimit(query.getLimit());
QueryResultList<Representation> results = performQuery(siteId, query);
List<JSONObject> jResultList = new ArrayList<JSONObject>(results.size());
// we need to know the highest score to normalise between [0..1]
double maxQueryScore = -1;
if (!results.isEmpty()) {
for (Representation r : results) {
if (maxQueryScore < 0) {
maxQueryScore = r.getFirst(resultScore.getUri(), Number.class).doubleValue();
}
JSONObject jResult = new JSONObject();
jResult.put("id", r.getId());
double similarity = 0.0;
// the name returned for the entity
String name = null;
for (Iterator<Text> labels = r.getText(NAME_FIELD); labels.hasNext(); ) {
Text label = labels.next();
if (label.getText().equalsIgnoreCase(rQuery.getQuery())) {
name = label.getText();
similarity = 1.0;
break;
}
double curSimilarity = Utils.levenshtein(rQuery.getQuery(), label.getText());
if (similarity < curSimilarity) {
name = label.getText();
similarity = curSimilarity;
}
}
// set the selected name
jResult.put("name", name);
Iterator<Reference> types = r.getReferences(TYPE_FIELD);
if (types != null && types.hasNext()) {
jResult.put("type", new JSONArray(ModelUtils.asCollection(types)));
}
double normalisedScore = r.getFirst(resultScore.getUri(), Number.class).doubleValue();
normalisedScore = normalisedScore * similarity / maxQueryScore;
jResult.put("score", normalisedScore);
jResult.put("match", similarity >= 0);
jResultList.add(jResult);
}
}
// else no results ... nothing todo
// sort results based on score
Collections.sort(jResultList, resultScoreComparator);
JSONObject jResultContainer = new JSONObject();
jResultContainer.put("result", new JSONArray(jResultList));
return jResultContainer;
}
use of org.apache.stanbol.entityhub.servicesapi.query.FieldQuery in project stanbol by apache.
the class JerseyUtils method createFieldQueryForFindRequest.
// /**
// * Returns the {@link FieldQuery} based on the JSON formatted String (in case
// * of "application/x-www-form-urlencoded" requests) or file (in case of
// * "multipart/form-data" requests).<p>
// * @param query the string containing the JSON serialised FieldQuery or
// * <code>null</code> in case of a "multipart/form-data" request
// * @param file the temporary file holding the data parsed by the request to
// * the web server in case of a "multipart/form-data" request or <code>null</code>
// * in case of the "application/x-www-form-urlencoded" request.
// * @return the FieldQuery parsed from the string provided by one of the two
// * parameters
// * @throws WebApplicationException if both parameter are <code>null</code> or
// * if the string provided by both parameters could not be used to parse a
// * {@link FieldQuery} instance.
// */
// public static FieldQuery parseFieldQuery(String query, File file) throws WebApplicationException {
// if(query == null && file == null) {
// throw new WebApplicationException(new IllegalArgumentException("Query Requests MUST define the \"query\" parameter"), Response.Status.BAD_REQUEST);
// }
// FieldQuery fieldQuery = null;
// JSONException exception = null;
// if(query != null){
// try {
// fieldQuery = JSONToFieldQuery.fromJSON(queryFactory,query);
// } catch (JSONException e) {
// log.warn("unable to parse FieldQuery from \"application/x-www-form-urlencoded\" encoded query string "+query,e);
// fieldQuery = null;
// exception = e;
// }
// } //else no query via application/x-www-form-urlencoded parsed
// if(fieldQuery == null && file != null){
// try {
// query = FileUtils.readFileToString(file);
// fieldQuery = JSONToFieldQuery.fromJSON(queryFactory,query);
// } catch (IOException e) {
// throw new WebApplicationException(e, Response.Status.INTERNAL_SERVER_ERROR);
// } catch (JSONException e) {
// log.warn("unable to parse FieldQuery from \"multipart/form-data\" encoded query string "+query,e);
// exception = e;
// }
// }//fieldquery already initialised or no query via multipart/form-data parsed
// if(fieldQuery == null){
// throw new WebApplicationException(new IllegalArgumentException("Unable to parse FieldQuery form the parsed query String:"+query, exception),Response.Status.BAD_REQUEST);
// }
// return fieldQuery;
// }
/**
* Creates an {@link FieldQuery} for parameters parsed by the /find requests
* supported by the /symbol, /sites and {siteId} RESTful endpoints.
* TODO: This has to be refactored to use "EntityQuery" as soon as Multiple
* query types are implemented for the Entityhub.
* @param name the name pattern to search entities for (required)
* @param field the field used to search for entities (required)
* @param language the language of the parsed name pattern (optional)
* @param limit the maximum number of result (optional)
* @param offset the offset of the first result (optional)
* @return the {@link FieldQuery} representing the parsed parameter
* @throws WebApplicationException in case the parsed name pattern is invalid.
* The validation of this required parameter provided by the Request is done
* by this method.
* @throws IllegalArgumentException in case the parsed field is invalid. Callers
* of this method need to ensure that this parameter is set to an valid value.
*/
public static FieldQuery createFieldQueryForFindRequest(String name, String field, String language, Integer limit, Integer offset, String ldpath) throws WebApplicationException, IllegalArgumentException {
if (name == null || name.trim().isEmpty()) {
// as an bad Requested sent by the user
throw new WebApplicationException(new IllegalArgumentException("The parsed \"name\" pattern to search entities for MUST NOT be NULL nor EMPTY"), Response.Status.BAD_REQUEST);
} else {
name = name.trim();
}
if (field == null || field.trim().isEmpty()) {
// valid data.
throw new IllegalArgumentException("The parsed search \"field\" MUST NOT be NULL nor EMPTY");
} else {
field = field.trim();
}
log.debug("Process Find Request:");
log.debug(" > name : " + name);
log.debug(" > field : " + field);
log.debug(" > lang : " + language);
log.debug(" > limit : " + limit);
log.debug(" > offset: " + offset);
log.debug(" > ldpath: " + ldpath);
FieldQuery query;
if (ldpath != null && !ldpath.isEmpty()) {
// STANBOL-417
query = new LDPathFieldQueryImpl();
((LDPathFieldQueryImpl) query).setLDPathSelect(ldpath);
} else {
// if no LDPath is parsed select the default field
query = queryFactory.createFieldQuery();
Collection<String> selectedFields = new ArrayList<String>();
// select also the field used to find entities
selectedFields.add(field);
query.addSelectedFields(selectedFields);
}
if (language == null || language.trim().isEmpty()) {
query.setConstraint(field, new TextConstraint(name, PatternType.wildcard, false));
} else {
query.setConstraint(field, new TextConstraint(name, PatternType.wildcard, false, language));
}
if (limit != null && limit > 0) {
query.setLimit(limit);
}
if (offset != null && offset > 0) {
query.setOffset(offset);
}
return query;
}
use of org.apache.stanbol.entityhub.servicesapi.query.FieldQuery in project stanbol by apache.
the class EntityhubRootResource method findEntity.
@POST
@Path("/find")
@Produces({ APPLICATION_JSON, RDF_XML, N3, TURTLE, X_TURTLE, RDF_JSON, N_TRIPLE, TEXT_HTML })
public Response findEntity(@FormParam(value = "name") String name, @FormParam(value = "field") String parsedField, @FormParam(value = "lang") String language, @FormParam(value = "limit") Integer limit, @FormParam(value = "offset") Integer offset, // solution!
@FormParam(value = "select") String select, @FormParam(value = "ldpath") String ldpath, @Context HttpHeaders headers) {
log.debug("/find Request");
final MediaType acceptedMediaType = getAcceptableMediaType(headers, ENTITY_SUPPORTED_MEDIA_TYPE_INCL_HTML, MediaType.APPLICATION_JSON_TYPE);
if (name == null || name.isEmpty()) {
if (acceptedMediaType.isCompatible(TEXT_HTML_TYPE)) {
// return HTML docu
ResponseBuilder rb = Response.ok(new Viewable("find", this));
rb.header(HttpHeaders.CONTENT_TYPE, TEXT_HTML + "; charset=utf-8");
// addCORSOrigin(servletContext, rb, headers);
return rb.build();
} else {
return Response.status(Status.BAD_REQUEST).entity("The name must not be null nor empty for find requests. Missing parameter name.\n").header(HttpHeaders.ACCEPT, acceptedMediaType).build();
}
} else {
final String property;
if (parsedField == null) {
property = DEFAULT_FIND_FIELD;
} else {
parsedField = parsedField.trim();
if (parsedField.isEmpty()) {
property = DEFAULT_FIND_FIELD;
} else {
property = nsPrefixService.getFullName(parsedField);
if (property == null) {
String messsage = String.format("The prefix '%s' of the parsed field '%' is not " + "mapped to any namespace. Please parse the full URI instead!\n", NamespaceMappingUtils.getPrefix(parsedField), parsedField);
return Response.status(Status.BAD_REQUEST).entity(messsage).header(HttpHeaders.ACCEPT, acceptedMediaType).build();
}
}
}
FieldQuery query = JerseyUtils.createFieldQueryForFindRequest(name, property, language, limit == null || limit < 1 ? DEFAULT_FIND_RESULT_LIMIT : limit, offset, ldpath);
// For the Entityhub we support to select additional fields for results
// of find requests. For the Sites and {site} endpoint this is currently
// deactivated because of very bad performance with OPTIONAL graph patterns
// in SPARQL queries.
Collection<String> additionalSelectedFields = new ArrayList<String>();
if (select != null && !select.isEmpty()) {
for (String selected : select.trim().split(" ")) {
if (selected != null && !selected.isEmpty()) {
additionalSelectedFields.add(selected);
}
}
}
query.addSelectedFields(additionalSelectedFields);
return executeQuery(query, headers, acceptedMediaType);
}
}
Aggregations