public void startScanFeatures(Observation obsr) {
currentFeatures.clear();
currentFeatureIdx = 0;
// scan over all context predicates
for (int i = 0; i < obsr.cps.length; i++) {
Element elem = (Element) dict.dict.get(new Integer(obsr.cps[i]));
if (elem == null) {
continue;
}
if (!(elem.isScanned)) {
// scan all labels for features
Iterator it = elem.lbCntFidxes.keySet().iterator();
while (it.hasNext()) {
Integer labelInt = (Integer) it.next();
CountFIdx cntFIdx = (CountFIdx) elem.lbCntFidxes.get(labelInt);
if (cntFIdx.fidx >= 0) {
Feature f = new Feature();
f.FeatureInit(labelInt.intValue(), obsr.cps[i]);
f.idx = cntFIdx.fidx;
elem.cpFeatures.add(f);
}
}
elem.isScanned = true;
}
for (int j = 0; j < elem.cpFeatures.size(); j++) {
currentFeatures.add(elem.cpFeatures.get(j));
}
}
}
public static void printNeighbors(ArrayList neighbors) {
int i = 0;
for(Neighbor neighbor : neighbors) {
Instance instance = neighbor.getInstance();
System.out.println("\nNeighbor " + (i + 1) + ", distance: " + neighbor.getDistance());
i++;
for(Feature f : instance.getAttributes()) {
System.out.print(f.getName() + ": ");
if(f instanceof Category) {
System.out.println(((Category)f).getCategory().toString());
}
else if(f instanceof Distance) {
System.out.println(((Distance)f).getDistance().toString());
}
else if (f instanceof Expiration) {
System.out.println(((Expiration)f).getExpiry().toString());
}
else if (f instanceof Handset) {
System.out.print(((Handset)f).getOs().toString() + ", ");
System.out.println(((Handset)f).getDevice().toString());
}
else if (f instanceof Offer) {
System.out.println(((Offer)f).getOfferType().toString());
}
else if (f instanceof WSAction) {
System.out.println(((WSAction)f).getAction().toString());
}
}
}
}
/*
* Create and return a new feature object with the given parameters.
*/
public Feature createFeature(Site site, String id, String version, String url) {
Feature result = new Feature(site);
result.setId(id);
result.setVersion(version);
result.setUrl(url);
return result;
}
public static void add(Object tag, Geometry geom, String msg) {
if (!Debug.isDebugging()) return;
FeatureDataset fd = getDebugFeatureDataset(tag);
Feature feature = new BasicFeature(fd.getFeatureSchema());
feature.setGeometry(geom);
feature.setAttribute(MESG_ATTR_NAME, msg);
fd.add(feature);
}
public static Feature createLineSegmentFeature(
FeatureSchema fs, Coordinate p0, Coordinate p1, String msg) {
Feature feature = new BasicFeature(fs);
LineString lineSeg = fact.createLineString(new Coordinate[] {p0, p1});
feature.setGeometry(lineSeg);
feature.setAttribute(MESG_ATTR_NAME, msg);
return feature;
}
private String getTypeOfResponseVariable(String responseVariable, List<Feature> features) {
String type = null;
for (Feature feature : features) {
if (feature.getName().equals(responseVariable)) {
type = feature.getType();
}
}
return type;
}
public void writeFeatures(PrintWriter fout) throws IOException {
// write the number of features
fout.println(Integer.toString(features.size()));
for (int i = 0; i < features.size(); i++) {
Feature f = (Feature) features.get(i);
fout.println(f.toString(data.cpInt2Str, data.lbInt2Str));
}
// wirte the line ###...
fout.println(Option.modelSeparator);
}
/**
* look at all the data in the column of the featurecollection, and find the largest string!
*
* @param fc features to look at
* @param attributeNumber which of the column to test.
*/
int findMaxStringLength(FeatureCollection fc, int attributeNumber) {
int l;
int maxlen = 0;
Feature f;
for (Iterator i = fc.iterator(); i.hasNext(); ) {
f = (Feature) i.next();
l = f.getString(attributeNumber).length();
if (l > maxlen) {
maxlen = l;
}
}
return maxlen;
}
public double classify(List<String> words) {
double score = 0.0;
double right = 1.0;
int count = 0;
List<Feature> features = kindTagging(words);
for (Feature feature : features) {
// logger.info("details: {}", feature.getText() + ": " + feature.getScore() + "; "
// + feature.getMultiple() + "; " + feature.getKind());
if (feature.getMultiple() != 0) {
right *= feature.getMultiple();
}
if (feature.getScore() != 0) {
score += right * feature.getScore();
count++;
right = 1.0;
}
if (feature.getMultiple() == 0) {
right = 1.0;
}
}
if (count != 0) {
score = score / count;
}
// System.out.println("sss " +score);
return score;
}
public static void printClassificationInstance(Instance classificationInstance) {
for(Feature f : classificationInstance.getAttributes()) {
System.out.print(f.getName() + ": ");
if(f instanceof Category) {
System.out.println(((Category)f).getCategory().toString());
}
else if(f instanceof Distance) {
System.out.println(((Distance)f).getDistance().toString());
}
else if (f instanceof Expiration) {
System.out.println(((Expiration)f).getExpiry().toString());
}
else if (f instanceof Handset) {
System.out.print(((Handset)f).getOs().toString() + ", ");
System.out.println(((Handset)f).getDevice().toString());
}
else if (f instanceof Offer) {
System.out.println(((Offer)f).getOfferType().toString());
}
else if (f instanceof WSAction) {
System.out.println(((WSAction)f).getAction().toString());
}
}
}
/**
* Add a feature to this map. Note that the start of the feature should be a smaller number than
* the stop of the feature, regardless of the strand. The only case where start is larger than the
* stop is when the feature runs across the start/stop boundary, for example 6899-10 on a 7000bp
* plasmid.
*
* @param type one of the following: origin_of_replication, promoter, terminator,
* selectable_marker, regulatory_sequence, tag, other_gene, reporter_gene,
* unique_restriction_site, restriction_site.
* @param name the name of the feature, such as EcoRI.
* @param start the start position of the feature. Must be between 1 and the length of the
* plasmid.
* @param stop the end position of the feature. Must be between 1 and the length of the plasmid.
* @param strand the strand of the feature. Can be Plasmid.FORWARD, Plasmid.REVERSE, or
* Plasmid.NO_DIRECTION.
*/
public void addFeature(String type, String name, int start, int stop, int strand) {
// add the feature to the plasmid.
int decoration;
int label;
Color color;
// String slot;
if (start > size) {
start = size;
}
if (start < 1) {
start = 1;
}
if (stop > size) {
stop = size;
}
if (stop < 1) {
stop = 1;
}
try {
color = getFeatureColor(type);
decoration = getFeatureDecoration(type, strand);
label = getLabelType();
addItemToLegend(type, strand);
} catch (NullPointerException e) {
color = new Color(0, 0, 128); // navy
if (colorScheme == REGULAR) {
color = new Color(0, 0, 128); // navy
} else if (colorScheme == INVERSE) {
color = new Color(0, 128, 128); // teal
}
decoration = DECORATION_STANDARD;
label = LABEL;
}
// create a feature and a feature range
// then figure out which feature slot to put the feature in.
Feature feature = new Feature(showShading);
FeatureRange featureRange = new FeatureRange(feature, start, stop);
featureRange.setDecoration(decoration);
featureRange.setColor(color);
featureRange.setOpacity(opacity);
featureRange.setShowLabel(label);
if ((showPositions)
&& ((type.equalsIgnoreCase("restriction_site"))
|| (type.equalsIgnoreCase("unique_restriction_site")))) {
featureRange.setLabel(name + " " + start);
} else {
if (addCategoryInfo) {
try {
featureRange.setLabel(name + (String) FEATURE_CATEGORIES.get(type));
} catch (NullPointerException e) {
featureRange.setLabel(name);
}
} else {
featureRange.setLabel(name);
}
}
if ((type.equalsIgnoreCase("restriction_site"))
|| (type.equalsIgnoreCase("unique_restriction_site"))) {
if (restrictionSlot == null) {
restrictionSlot = new FeatureSlot(DIRECT_STRAND, showShading);
restrictionSlot.setFeatureThickness(1f);
}
feature.setFeatureSlot(restrictionSlot);
} else if ((strand == NO_DIRECTION) || (strand == FORWARD)) {
if (forwardSlot0 == null) {
forwardSlot0 = new FeatureSlot(DIRECT_STRAND, showShading);
feature.setFeatureSlot(forwardSlot0);
} else if (forwardSlot0.isRoom(feature)) {
feature.setFeatureSlot(forwardSlot0);
} else if (forwardSlot1 == null) {
forwardSlot1 = new FeatureSlot(DIRECT_STRAND, showShading);
feature.setFeatureSlot(forwardSlot1);
} else if (forwardSlot1.isRoom(feature)) {
feature.setFeatureSlot(forwardSlot1);
} else if (forwardSlot2 == null) {
forwardSlot2 = new FeatureSlot(DIRECT_STRAND, showShading);
feature.setFeatureSlot(forwardSlot2);
} else if (forwardSlot2.isRoom(feature)) {
feature.setFeatureSlot(forwardSlot2);
} else if (forwardSlot3 == null) {
forwardSlot3 = new FeatureSlot(DIRECT_STRAND, showShading);
feature.setFeatureSlot(forwardSlot3);
} else if (forwardSlot3.isRoom(feature)) {
feature.setFeatureSlot(forwardSlot3);
} else if (forwardSlot4 == null) {
forwardSlot4 = new FeatureSlot(DIRECT_STRAND, showShading);
feature.setFeatureSlot(forwardSlot4);
} else if (forwardSlot4.isRoom(feature)) {
feature.setFeatureSlot(forwardSlot4);
} else if (forwardSlot5 == null) {
forwardSlot5 = new FeatureSlot(DIRECT_STRAND, showShading);
feature.setFeatureSlot(forwardSlot5);
} else if (forwardSlot5.isRoom(feature)) {
feature.setFeatureSlot(forwardSlot5);
} else if (forwardSlot6 == null) {
forwardSlot6 = new FeatureSlot(DIRECT_STRAND, showShading);
feature.setFeatureSlot(forwardSlot6);
} else if (forwardSlot6.isRoom(feature)) {
feature.setFeatureSlot(forwardSlot6);
} else if (forwardSlot7 == null) {
forwardSlot7 = new FeatureSlot(DIRECT_STRAND, showShading);
feature.setFeatureSlot(forwardSlot7);
} else {
feature.setFeatureSlot(forwardSlot7);
}
} else if (strand == REVERSE) {
if (reverseSlot0 == null) {
reverseSlot0 = new FeatureSlot(REVERSE_STRAND, showShading);
feature.setFeatureSlot(reverseSlot0);
} else if (reverseSlot0.isRoom(feature)) {
feature.setFeatureSlot(reverseSlot0);
} else if (reverseSlot1 == null) {
reverseSlot1 = new FeatureSlot(REVERSE_STRAND, showShading);
feature.setFeatureSlot(reverseSlot1);
} else if (reverseSlot1.isRoom(feature)) {
feature.setFeatureSlot(reverseSlot1);
} else if (reverseSlot2 == null) {
reverseSlot2 = new FeatureSlot(REVERSE_STRAND, showShading);
feature.setFeatureSlot(reverseSlot2);
} else if (reverseSlot2.isRoom(feature)) {
feature.setFeatureSlot(reverseSlot2);
} else if (reverseSlot3 == null) {
reverseSlot3 = new FeatureSlot(REVERSE_STRAND, showShading);
feature.setFeatureSlot(reverseSlot3);
} else if (reverseSlot3.isRoom(feature)) {
feature.setFeatureSlot(reverseSlot3);
} else if (reverseSlot4 == null) {
reverseSlot4 = new FeatureSlot(REVERSE_STRAND, showShading);
feature.setFeatureSlot(reverseSlot4);
} else if (reverseSlot4.isRoom(feature)) {
feature.setFeatureSlot(reverseSlot4);
} else if (reverseSlot5 == null) {
reverseSlot5 = new FeatureSlot(REVERSE_STRAND, showShading);
feature.setFeatureSlot(reverseSlot5);
} else if (reverseSlot5.isRoom(feature)) {
feature.setFeatureSlot(reverseSlot5);
} else if (reverseSlot6 == null) {
reverseSlot6 = new FeatureSlot(REVERSE_STRAND, showShading);
feature.setFeatureSlot(reverseSlot6);
} else if (reverseSlot6.isRoom(feature)) {
feature.setFeatureSlot(reverseSlot6);
} else if (reverseSlot7 == null) {
reverseSlot7 = new FeatureSlot(REVERSE_STRAND, showShading);
feature.setFeatureSlot(reverseSlot7);
} else {
feature.setFeatureSlot(reverseSlot7);
}
}
}
public List<Feature> kindTagging(List<String> words) {
List<Feature> features = new ArrayList<Feature>();
for (String word : words) {
Feature feature = new Feature();
feature.setText(word);
if (isPrivativeWord(word)) {
// System.out.println("bubnububub");
feature.setKind(Enum.valueOf(WordType.class, "PRIVATIVE"));
feature.setMultiple(-1.0);
} else if (isDecoVeryWord(word)) {
feature.setKind(Enum.valueOf(WordType.class, "DECORATEVERY"));
feature.setMultiple(2.0);
} else if (isDecoLittleWord(word)) {
feature.setKind(Enum.valueOf(WordType.class, "DECORATELITTLE"));
feature.setMultiple(0.5);
} else if (isPositiveWord(word)) {
feature.setKind(Enum.valueOf(WordType.class, "POSITIVE"));
feature.setMultiple(1.0);
feature.setScore(posiLexicon.get(word));
} else if (isNegativeWord(word)) {
feature.setKind(Enum.valueOf(WordType.class, "NEGATIVE"));
feature.setMultiple(-1.0);
feature.setScore(negaLexicon.get(word));
}
features.add(feature);
}
return features;
}
// generating features
public void generateFeatures() {
if (features != null) {
features.clear();
} else {
features = new ArrayList();
}
if (fmap != null) {
fmap.clear();
} else {
fmap = new HashMap();
}
if (currentFeatures != null) {
currentFeatures.clear();
} else {
currentFeatures = new ArrayList();
}
if (data.trnData == null || dict.dict == null) {
System.out.println("No data or dictionary for generating features");
return;
}
// scan over data list
for (int i = 0; i < data.trnData.size(); i++) {
Observation obsr = (Observation) data.trnData.get(i);
for (int j = 0; j < obsr.cps.length; j++) {
Element elem;
CountFIdx cntFIdx;
elem = (Element) dict.dict.get(new Integer(obsr.cps[j]));
if (elem != null) {
if (elem.count <= option.cpRareThreshold) {
// skip this context predicate, it is too rare
continue;
}
cntFIdx = (CountFIdx) elem.lbCntFidxes.get(new Integer(obsr.humanLabel));
if (cntFIdx != null) {
if (cntFIdx.count <= option.fRareThreshold) {
// skip this feature, it is too rare
continue;
}
} else {
// not found in the dictionary, then skip
continue;
}
} else {
// not found in the dictionary, then skip
continue;
}
// update the feature
Feature f = new Feature(obsr.humanLabel, obsr.cps[j]);
f.strId2Idx(fmap);
if (f.idx < 0) {
// new feature, add to the feature list
addFeature(f);
// update the feature index in the dictionary
cntFIdx.fidx = f.idx;
elem.chosen = 1;
}
}
}
option.numFeatures = features.size();
}
// adding a feature
public void addFeature(Feature f) {
f.strId2IdxAdd(fmap);
features.add(f);
}
/**
* Write a dbf file with the information from the featureCollection.
*
* @param featureCollection column data from collection
* @param fname name of the dbf file to write to
*/
void writeDbf(FeatureCollection featureCollection, String fname) throws Exception {
DbfFileWriter dbf;
FeatureSchema fs;
int t;
int f;
int u;
int num;
fs = featureCollection.getFeatureSchema();
// -1 because one of the columns is geometry
DbfFieldDef[] fields = new DbfFieldDef[fs.getAttributeCount() - 1];
// dbf column type and size
f = 0;
for (t = 0; t < fs.getAttributeCount(); t++) {
AttributeType columnType = fs.getAttributeType(t);
String columnName = fs.getAttributeName(t);
if (columnType == AttributeType.INTEGER) {
fields[f] = new DbfFieldDef(columnName, 'N', 16, 0);
f++;
} else if (columnType == AttributeType.DOUBLE) {
fields[f] = new DbfFieldDef(columnName, 'N', 33, 16);
f++;
} else if (columnType == AttributeType.STRING) {
int maxlength = findMaxStringLength(featureCollection, t);
if (maxlength > 255) {
throw new Exception(
"ShapefileWriter does not support strings longer than 255 characters");
}
fields[f] = new DbfFieldDef(columnName, 'C', maxlength, 0);
f++;
} else if (columnType == AttributeType.DATE) {
fields[f] = new DbfFieldDef(columnName, 'D', 8, 0);
f++;
} else if (columnType == AttributeType.GEOMETRY) {
// do nothing - the .shp file handles this
} else {
throw new Exception("Shapewriter: unsupported AttributeType found in featurecollection.");
}
}
// write header
dbf = new DbfFileWriter(fname);
dbf.writeHeader(fields, featureCollection.size());
// write rows
num = featureCollection.size();
List features = featureCollection.getFeatures();
for (t = 0; t < num; t++) {
// System.out.println("dbf: record "+t);
Feature feature = (Feature) features.get(t);
Vector DBFrow = new Vector();
// make data for each column in this feature (row)
for (u = 0; u < fs.getAttributeCount(); u++) {
AttributeType columnType = fs.getAttributeType(u);
if (columnType == AttributeType.INTEGER) {
Object a = feature.getAttribute(u);
if (a == null) {
DBFrow.add(new Integer(0));
} else {
DBFrow.add((Integer) a);
}
} else if (columnType == AttributeType.DOUBLE) {
Object a = feature.getAttribute(u);
if (a == null) {
DBFrow.add(new Double(0.0));
} else {
DBFrow.add((Double) a);
}
} else if (columnType == AttributeType.DATE) {
Object a = feature.getAttribute(u);
if (a == null) {
DBFrow.add("");
} else {
DBFrow.add(DbfFile.DATE_PARSER.format((Date) a));
}
} else if (columnType == AttributeType.STRING) {
Object a = feature.getAttribute(u);
if (a == null) {
DBFrow.add(new String(""));
} else {
// MD 16 jan 03 - added some defensive programming
if (a instanceof String) {
DBFrow.add(a);
} else {
DBFrow.add(a.toString());
}
}
}
}
dbf.writeRecord(DBFrow);
}
dbf.close();
}
public void addDefinition(Feature definition) {
definitions.put(definition.getID(), definition);
}
public String getDescriptionForID(String id) {
Feature x = (Feature) definitions.get(id);
if (x != null) return x.getDescription();
else return "?";
}
public Feature buildFeature(String id) {
Feature x = (Feature) definitions.get(id);
if (x != null) return (Feature) x.clone();
Debug.byebye("Feature " + id + " not found");
return null;
}
