List of usage examples for org.apache.lucene.document Field stringValue
@Override
public String stringValue()
From source file:at.lux.retrieval.vectorspace.ElementTextVectorSimilarityTest.java
License:Open Source License
public void testSimilarity() throws IOException, JDOMException { ElementTextVectorSimilarity sim = new ElementTextVectorSimilarity(); double distance = sim.getSimilarity(d1, d1); System.out.println("distance = " + distance); distance = sim.getSimilarity(d1, d2); System.out.println("distance = " + distance); distance = sim.getSimilarity(d2, d1); System.out.println("distance = " + distance); IndexReader reader = IndexReader.open("testdata/idx_paths"); System.out.println("Loading documents and adding them to corpus ..."); for (int i = 0; i < reader.numDocs(); i++) { // Graph g_idx = new Graph(reader.document(i).getField("graph").stringValue()); Field[] files = reader.document(i).getFields("file"); for (Field file : files) { Document d = saxBuilder.build(file.stringValue()); sim.addToCorpus(d);// ww w . j a v a 2 s. com } } System.out.println(""); distance = sim.getSimilarity(d1, d1, ElementTextVectorSimilarity.WeightType.TfIdf); System.out.println("distance = " + distance); distance = sim.getSimilarity(d1, d2, ElementTextVectorSimilarity.WeightType.TfIdf); System.out.println("distance = " + distance); distance = sim.getSimilarity(d2, d1, ElementTextVectorSimilarity.WeightType.TfIdf); System.out.println("distance = " + distance); distance = sim.getSimilarity(d2, d2, ElementTextVectorSimilarity.WeightType.TfIdf); System.out.println("distance = " + distance); System.out.println(""); distance = sim.getSimilarity(d1, d1, ElementTextVectorSimilarity.WeightType.BM25); System.out.println("distance = " + distance); distance = sim.getSimilarity(d1, d2, ElementTextVectorSimilarity.WeightType.BM25); System.out.println("distance = " + distance); distance = sim.getSimilarity(d2, d1, ElementTextVectorSimilarity.WeightType.BM25); System.out.println("distance = " + distance); distance = sim.getSimilarity(d2, d2, ElementTextVectorSimilarity.WeightType.BM25); System.out.println("distance = " + distance); }
From source file:axiom.db.utils.LuceneManipulator.java
License:Open Source License
protected Document convertDocument(Document doc) { Document ndoc = new Document(); Enumeration e = doc.fields(); while (e.hasMoreElements()) { Field f = (Field) e.nextElement(); Field.Store currstore = Field.Store.YES; if (!f.isStored()) { currstore = Field.Store.NO; } else if (f.isCompressed()) { currstore = Field.Store.COMPRESS; }/*w ww . j av a 2s . com*/ Field.Index curridx = Field.Index.UN_TOKENIZED; if (!f.isIndexed()) { curridx = Field.Index.NO; } else if (f.isTokenized()) { curridx = Field.Index.TOKENIZED; } String name = f.name(); String value = f.stringValue(); ndoc.add(new Field(name, value, currstore, curridx)); } return ndoc; }
From source file:axiom.objectmodel.dom.convert.LuceneVersion2Convertor.java
License:Open Source License
public Document convertDocument(Document doc) { Document ndoc = new Document(); Enumeration e = doc.fields(); while (e.hasMoreElements()) { Field f = (Field) e.nextElement(); Field.Store currstore = Field.Store.YES; if (!f.isStored()) { currstore = Field.Store.NO; } else if (f.isCompressed()) { currstore = Field.Store.COMPRESS; }/*from w w w .j a v a 2s . c o m*/ Field.Index curridx = Field.Index.UN_TOKENIZED; if (!f.isIndexed()) { curridx = Field.Index.NO; } else if (f.isTokenized()) { curridx = Field.Index.TOKENIZED; } String name = f.name(); String value = f.stringValue(); if (name.startsWith("__")) { name = name.substring(1); if ("_state".equals(name)) { name = "_status"; } else if ("_prototype".equals(name) || "_parentproto".equals(name)) { if ("HopObject".equalsIgnoreCase(value)) { value = "AxiomObject"; } } } ndoc.add(new Field(name, value, currstore, curridx)); } return ndoc; }
From source file:axiom.objectmodel.dom.convert.LuceneVersion3Convertor.java
License:Open Source License
public Document convertDocument(Document doc) { Document ndoc = new Document(); Enumeration e = doc.fields(); while (e.hasMoreElements()) { Field f = (Field) e.nextElement(); Field.Store currstore = Field.Store.YES; if (!f.isStored()) { currstore = Field.Store.NO; } else if (f.isCompressed()) { currstore = Field.Store.COMPRESS; }/* ww w .j ava 2s. co m*/ Field.Index curridx = Field.Index.UN_TOKENIZED; if (!f.isIndexed()) { curridx = Field.Index.NO; } else if (f.isTokenized()) { curridx = Field.Index.TOKENIZED; } String name = f.name(); String value = f.stringValue(); if (name.startsWith("__")) { name = name.substring(1); if ("_state".equals(name)) { name = "_status"; } else if ("_prototype".equals(name) || "_parentproto".equals(name)) { if ("HopObject".equalsIgnoreCase(value)) { value = "AxiomObject"; } } } ndoc.add(new Field(name, value, currstore, curridx)); } return ndoc; }
From source file:axiom.objectmodel.dom.convert.LuceneVersion4Convertor.java
License:Open Source License
public Document convertDocument(Document doc) { Document ndoc = new Document(); Enumeration e = doc.fields(); final String prototype = doc.get("_prototype"); final Prototype proto = app.getPrototypeByName(prototype); while (e.hasMoreElements()) { Field f = (Field) e.nextElement(); Field.Store currstore = Field.Store.YES; if (!f.isStored()) { currstore = Field.Store.NO; } else if (f.isCompressed()) { currstore = Field.Store.COMPRESS; }//from w w w . j a v a2 s .co m Field.Index curridx = Field.Index.UN_TOKENIZED; if (!f.isIndexed()) { curridx = Field.Index.NO; } else if (f.isTokenized()) { curridx = Field.Index.TOKENIZED; } String name = f.name(); String value = f.stringValue(); final int type = this.getTypeForProperty(proto, name); if (type == IProperty.DATE) { value = formatDate(value); } else if (type == IProperty.TIME) { value = formatTime(value); } ndoc.add(new Field(name, value, currstore, curridx)); } return ndoc; }
From source file:axiom.objectmodel.dom.convert.LuceneVersion5Convertor.java
License:Open Source License
protected Document convertDocument(Document doc) { Document ndoc = new Document(); Enumeration e = doc.fields(); if (doc.get("cms_sortable_prototype") == null) { ndoc.add(new Field("cms_sortable_prototype", getDisplayPrototype(cms_props, doc.get("_prototype")), Field.Store.YES, Field.Index.UN_TOKENIZED)); }/*from w w w . java 2 s . co m*/ if (doc.get("cms_sortabletitle") == null && doc.get("title") != null) { ndoc.add(new Field("cms_sortabletitle", doc.get("title").toLowerCase(), Field.Store.YES, Field.Index.UN_TOKENIZED)); } while (e.hasMoreElements()) { Field f = (Field) e.nextElement(); Field.Store currstore = Field.Store.YES; if (!f.isStored()) { currstore = Field.Store.NO; } else if (f.isCompressed()) { currstore = Field.Store.COMPRESS; } Field.Index curridx = Field.Index.UN_TOKENIZED; if (!f.isIndexed()) { curridx = Field.Index.NO; } else if (f.isTokenized()) { curridx = Field.Index.TOKENIZED; } String name = f.name(); String value = f.stringValue(); ndoc.add(new Field(name, value, currstore, curridx)); } return ndoc; }
From source file:axiom.objectmodel.dom.convert.LuceneVersion6Convertor.java
License:Open Source License
protected Document convertDocument(Document doc) { Document ndoc = new Document(); Enumeration e = doc.fields(); while (e.hasMoreElements()) { Field f = (Field) e.nextElement(); Field.Store currstore = Field.Store.YES; if (!f.isStored()) { currstore = Field.Store.NO; } else if (f.isCompressed()) { currstore = Field.Store.COMPRESS; }//from www .j ava 2s . com Field.Index curridx = Field.Index.UN_TOKENIZED; if (!f.isIndexed()) { curridx = Field.Index.NO; } else if (f.isTokenized()) { curridx = Field.Index.TOKENIZED; } String name = f.name(); String value = f.stringValue(); if (name.equals("cms_lasteditedby")) { Date last = new Date(Long.parseLong(doc.get("_lastmodified"))); String user = doc.get("lastmodifiedby"); if (user == null) { user = ""; } value = last.getDate() + " " + months[last.getMonth()] + " " + (last.getYear() + "").substring(1) + ", " + last.getHours() + ':' + (last.getMinutes() < 10 ? '0' + last.getMinutes() : last.getMinutes()) + " by " + user; } ndoc.add(new Field(name, value, currstore, curridx)); } return ndoc; }
From source file:axiom.objectmodel.dom.convert.LuceneVersion7Convertor.java
License:Open Source License
protected Document convertDocument(Document doc) throws Exception { Document ndoc = new Document(); Enumeration e = doc.fields(); final String prototype = doc.get("_prototype"); final Prototype proto = app.getPrototypeByName(prototype); while (e.hasMoreElements()) { Field f = (Field) e.nextElement(); Field.Store currstore = Field.Store.YES; if (!f.isStored()) { currstore = Field.Store.NO; } else if (f.isCompressed()) { currstore = Field.Store.COMPRESS; }//from www . j a va 2s.c om Field.Index curridx = Field.Index.UN_TOKENIZED; if (!f.isIndexed()) { curridx = Field.Index.NO; } else if (f.isTokenized()) { curridx = Field.Index.TOKENIZED; } String name = f.name(); String value = f.stringValue(); final int type = this.getTypeForProperty(proto, name); if (type == IProperty.DATE) { value = formatDate(value); } else if (type == IProperty.TIME) { value = formatTime(value); } else if (type == IProperty.TIMESTAMP) { value = formatTimestamp(value); } else if (type == IProperty.INTEGER) { value = formatInt(value); } else if (type == IProperty.FLOAT) { value = formatFloat(value); } else if (type == IProperty.SMALLFLOAT) { value = formatSmallFloat(value); } else if (type == IProperty.SMALLINT) { value = formatSmallInt(value); } ndoc.add(new Field(name, value, currstore, curridx)); } return ndoc; }
From source file:axiom.objectmodel.dom.convert.LuceneVersion8Convertor.java
License:Open Source License
public Document convertDocument(Document doc) { Document ndoc = new Document(); Enumeration e = doc.fields(); while (e.hasMoreElements()) { Field f = (Field) e.nextElement(); Field.Store currstore = Field.Store.YES; if (!f.isStored()) { currstore = Field.Store.NO; } else if (f.isCompressed()) { currstore = Field.Store.COMPRESS; }/*ww w.j ava2s . c om*/ Field.Index curridx = Field.Index.UN_TOKENIZED; if (!f.isIndexed()) { curridx = Field.Index.NO; } else if (f.isTokenized()) { curridx = Field.Index.TOKENIZED; } String name = f.name(); String value = f.stringValue(); ndoc.add(new Field(name, value, currstore, curridx)); } return ndoc; }
From source file:axiom.objectmodel.dom.convert.LuceneVersion9Convertor.java
License:Open Source License
public Document convertDocument(Document doc) { Field protoField = doc.getField(LuceneManager.PROTOTYPE); if (protoField != null && ("CMSTask".equalsIgnoreCase(protoField.stringValue()) || "CMSTaskContainer".equalsIgnoreCase(protoField.stringValue()))) { return null; }/*from w w w.j av a 2s . co m*/ Document ndoc = new Document(); Enumeration e = doc.fields(); String id = null, layer = null; Field idField = doc.getField(LuceneManager.ID); Field layerField = doc.getField(LuceneManager.LAYER_OF_SAVE); if (idField != null && layerField != null) { id = idField.stringValue(); layer = layerField.stringValue(); } while (e.hasMoreElements()) { Field f = (Field) e.nextElement(); Field.Store currstore = Field.Store.YES; if (!f.isStored()) { currstore = Field.Store.NO; } else if (f.isCompressed()) { currstore = Field.Store.COMPRESS; } Field.Index curridx = Field.Index.UN_TOKENIZED; if (!f.isIndexed()) { curridx = Field.Index.NO; } else if (f.isTokenized()) { curridx = Field.Index.TOKENIZED; } String name = f.name(); String value = f.stringValue(); if (!("84".equals(id) && "1".equals(layer) && "_task".equals(name)) && !("71".equals(id) && "1".equals(layer) && "_task".equals(name))) { ndoc.add(new Field(name, value, currstore, curridx)); } } return ndoc; }