I already have a CRF trained model that I have trained using SimpleTagger.
SimpleTagger.main(new String[] {
"--train", "true",
"--model-file", "/Desktop/crfmodel",
"--threads", "8",
"--training-proportion", "0.8",
"--weights", "dense",
"--test", "lab",
// "--orders", "2",
"/Desktop/annotations.txt"
});
I am planning to load this model and use it for tagging. I am using this code.
public static void main(String[] args) throws Exception {
//DOCS http://mallet.cs.umass.edu/classifier-devel.php
Instance instance = getMyInstance();
Classifier classifier = loadClassifier(Paths.get("/Desktop/crfmodel").toFile());
Labeling labeling = classifier.classify(instance).getLabeling();
Label l = labeling.getBestLabel();
System.out.print(instance);
System.out.println(l);
}
private static Classifier loadClassifier(File serializedFile)
throws FileNotFoundException, IOException, ClassNotFoundException {
ObjectInputStream ois = new ObjectInputStream (new FileInputStream(serializedFile));
Classifier crf = (Classifier) ois.readObject();
ois.close();
return crf;
}
When I try to do the above I get the following error
Exception in thread "main" java.lang.ClassCastException: cc.mallet.fst.CRF cannot be cast to cc.mallet.classify.Classifier
at TagClassifier.loadClassifier(TagClassifier.java:77)
at TagClassifier.main(TagClassifier.java:64)
The error is happening in line
Classifier crf = (Classifier) ois.readObject();
May I know why this is happening. Also, if there is a correct documented way to label an input using a trained model, can you please share any links/documentation? Thank you very much in advance!!!
I think I figured it out by looking at SimpleTagger code.
crfModel = loadClassifier(Paths.get("/Desktop/crfmodel").toFile());
pipe = crfModel.getInputPipe();
pipe.setTargetProcessing(false);
String formatted = getFormattedQuery(q);
Instance instance = pipe.pipe(new Instance(formatted, null, null, null));
Sequence sequence = (Sequence) instance.getData();
Sequence[] tags = tag(sequence, 3);
private static Sequence[] tag(Sequence input, int bestK) {
Sequence[] answers;
if (bestK == 1) {
answers = new Sequence[1];
answers[0] = crfModel.transduce(input);
} else {
MaxLatticeDefault lattice = new MaxLatticeDefault(crfModel, input, null);
answers = lattice.bestOutputSequences(bestK).toArray(new Sequence[0]);
}
return answers;
}
They're different things, so you can't cast one to the other. A CRF infers classes for each element in a sequence, so its output is an array of labels. A classifier takes one input and returns one label.
Related
I'm trying to create an application which takes in the smartphone sensor data (accelerometer and gyroscope) that predicts the driving style currently produced (eg. Right turn, Left turn ,etc). I've already converted the LSTM model into TFLite model and have imported it into Android Studio and the sample code below was created . What I understand is we need to create a byteBuffer to load the input into the model. Here is the sample code after I have imported it into the project:
try {
FullDataDriverprofiler model = FullDataDriverprofiler.newInstance(context);
// Creates inputs for reference.
TensorBuffer inputFeature0 = TensorBuffer.createFixedSize(new int[]{1, 50, 6}, DataType.FLOAT32);
inputFeature0.loadBuffer(byteBuffer);
// Runs model inference and gets result.
FullDataDriverprofiler.Outputs outputs = model.process(inputFeature0);
TensorBuffer outputFeature0 = outputs.getOutputFeature0AsTensorBuffer();
// Releases model resources if no longer used.
model.close();
} catch (IOException e) {
// TODO Handle the exception
}
I have initialized the sensors, and have stored them inside an ArrayList (I've stored the accelerometer and gyroscope data into an ArrayList where each axis takes in 50 values at a time):
private void predictActivities() {
//create new list to combine all arrayLists into one big list. Data will be our input variable
List<Float> data = new ArrayList<>();
if( ax.size() >= TIME_STAMP && ay.size() >= TIME_STAMP && az.size() >= TIME_STAMP
&& gx.size() >= TIME_STAMP && gy.size() >= TIME_STAMP && gz.size() >= TIME_STAMP)
{
data.addAll(ax.subList(0, TIME_STAMP));
data.addAll(ay.subList(0, TIME_STAMP));
data.addAll(az.subList(0, TIME_STAMP));
data.addAll(gx.subList(0, TIME_STAMP));
data.addAll(gy.subList(0, TIME_STAMP));
data.addAll(gz.subList(0, TIME_STAMP));
}
Log.d(TAG, "predictActivities: Data in List ArrayList"+ data);
I then convert the ArrayList into a Float array:
private float[] toFloatArray(List<Float> data){
int i = 0;
float[] array = new float[data.size()];
for (Float f: data){
array[i++] = (f !=null ? f: Float.NaN);
}
return array;
}
How do we feed the float array as input for the TFlite model? I've seen examples where they use TensorImage() to pass in the inputs but that is solely for images right? Thanks for the help!
I am identifying qualifications in a large corpus. I am using NamedEntityTagAnnotation.
Problem:
My annotations are read in as case sensitive. I want them to be case insensitive.
Hence
Bachelor's Degree DEGREE
does not need an additional entry of
Bachelor's degree DEGREE
I know this is possible. RegexNERAnnotator has a field for ignoreCase. But I don't know how to access RegexNERAnnotator through the API.
My current code (which I cadged off the internet and works apart from the case issue) is as follows:
String prevNeToken = "O";
String currNeToken = "O";
boolean newToken = true;
for (CoreLabel token : sentence.get(TokensAnnotation.class))
{
currNeToken = token.get(NamedEntityTagAnnotation.class);
String word = token.get(TextAnnotation.class);
if (currNeToken.equals("O"))
{
if (!prevNeToken.equals("O") && (sbuilder.length() > 0))
{
handleEntity(prevNeToken, sbuilder, tokens);
newToken = true;
}
continue;
}
if (newToken)
{
prevNeToken = currNeToken;
newToken = false;
sbuilder.append(word);
continue;
}
if (currNeToken.equals(prevNeToken))
{
sbuilder.append(" " + word);
}
else
{
handleEntity(prevNeToken, sbuilder, tokens);
newToken = true;
}
prevNeToken = currNeToken;
}
Any assistance would be greatly appreciated.
The answer is in how you set up the pipeline.
Properties props = new Properties();
props.put("annotators", "tokenize, ssplit, pos, lemma, ner, parse, regexner, depparse, natlog, openie");
//props.put("regexner.mapping", namedEntityPropertiesPath);
pipeline = new StanfordCoreNLP(props);
pipeline.addAnnotator(new TokensRegexNERAnnotator(namedEntityPropertiesPath, true));
Do not use props.put("regexner.mapping", namedEntityPropertiesPath);
Use pipeline.addAnnotator.
The first argument to the constructor is the path to your NER data file. The second is a boolean caseInsensitive.
Note, that this then uses Stanford's NER lists as well as your own. It also uses a more complex NER data file.
See http://nlp.stanford.edu/nlp/javadoc/javanlp/edu/stanford/nlp/pipeline/TokensRegexNERAnnotator.html
I trained OpenNLP NER-Model to detect a new Entity but when I am using this model I encountered the following Exception:
Exception in thread "main" java.lang.IllegalArgumentException:
opennlp.tools.util.InvalidFormatException:
Model version 1.6.0 is not supported by this (1.5.3) version of OpenNLP!
I am using OpenNLP version 1.6.0 and my source code is this:
String [] sentences = Fragmentation.getSentences(Document);
InputStream modelIn = new FileInputStream("Models/en-ner-cvskill.bin");
TokenNameFinderModel model = new TokenNameFinderModel(modelIn);
NameFinderME nameFinder = new NameFinderME(model);
String[] tokens = null;
Span nameSpans[] = null;
int i=0;
for (String sentence : sentences) {
tokens = null;
nameSpans = null;
System.out.println("Sentences: "+(++i)+"\n" + sentence);
tokens = Fragmentation.getTokens(sentence);
for(String token: tokens){
System.out.println("Token:-------------------: "+token);
}
nameSpans = nameFinder.find(tokens);
String SkillName = Arrays.toString(Span.spansToStrings(nameSpans, tokens));
for(Span name:nameSpans){
System.out.println("Skills: "+ name.toString());
}
System.out.println("Names-------------------:"+SkillName);
}
nameFinder.clearAdaptiveData();
Anyone please help me solve this problem..
I have finde out the problem..
Actually I was training namefinder of opennlp 1.6.0 and was using within the same version which is not possible with the current version(1.6.0) of the opennlp.
Now I trained the model of opennlp 1.5.3 and is using with opennlp 1.6.0 which is working fine!
I am trying to find whether a sentence is Positive or Negative in the following steps:
1.) Retrieving the Parts of speech(verbs, nouns, adjectives etc) from the sentence using the Stanford NLP parser.
2.) Using the SentiWordNet to find the Positive and Negative values related to each Part of Speech.
3.) Summing up the Positive and Negative values obtained to calculate a Net Positive and Net Negative value related to a sentence.
But the problem is that, the SentiWordNet return a list of Positive/Negative values based on different senses/contexts. Is it possible to pass a particular sentence along with the part of speech to the SentiWordNet parser, so that it can judge the sense/context automatically and returns only one pair of Positive and Negative value?
Or is there any other alternate solution to this problem?
Thanks.
SentoWordNet Demo Code
This may help you.
// Copyright 2013 Petter Törnberg
//
// This demo code has been kindly provided by Petter Törnberg <pettert#chalmers.se>
// for the SentiWordNet website.
//
// This program is free software: you can redistribute it and/or modify
// it under the terms of the GNU General Public License as published by
// the Free Software Foundation, either version 3 of the License, or
// (at your option) any later version.
//
// This program is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU General Public License for more details.
//
// You should have received a copy of the GNU General Public License
// along with this program. If not, see <http://www.gnu.org/licenses/>.
import java.io.BufferedReader;
import java.io.FileReader;
import java.io.IOException;
import java.util.HashMap;
import java.util.Map;
public class SentiWordNetDemoCode {
private Map<String, Double> dictionary;
public SentiWordNetDemoCode(String pathToSWN) throws IOException {
// This is our main dictionary representation
dictionary = new HashMap<String, Double>();
// From String to list of doubles.
HashMap<String, HashMap<Integer, Double>> tempDictionary = new HashMap<String, HashMap<Integer, Double>>();
BufferedReader csv = null;
try {
csv = new BufferedReader(new FileReader(pathToSWN));
int lineNumber = 0;
String line;
while ((line = csv.readLine()) != null) {
lineNumber++;
// If it's a comment, skip this line.
if (!line.trim().startsWith("#")) {
// We use tab separation
String[] data = line.split("\t");
String wordTypeMarker = data[0];
// Example line:
// POS ID PosS NegS SynsetTerm#sensenumber Desc
// a 00009618 0.5 0.25 spartan#4 austere#3 ascetical#2
// ascetic#2 practicing great self-denial;...etc
// Is it a valid line? Otherwise, through exception.
if (data.length != 6) {
throw new IllegalArgumentException(
"Incorrect tabulation format in file, line: "
+ lineNumber);
}
// Calculate synset score as score = PosS - NegS
Double synsetScore = Double.parseDouble(data[2])
- Double.parseDouble(data[3]);
// Get all Synset terms
String[] synTermsSplit = data[4].split(" ");
// Go through all terms of current synset.
for (String synTermSplit : synTermsSplit) {
// Get synterm and synterm rank
String[] synTermAndRank = synTermSplit.split("#");
String synTerm = synTermAndRank[0] + "#"
+ wordTypeMarker;
int synTermRank = Integer.parseInt(synTermAndRank[1]);
// What we get here is a map of the type:
// term -> {score of synset#1, score of synset#2...}
// Add map to term if it doesn't have one
if (!tempDictionary.containsKey(synTerm)) {
tempDictionary.put(synTerm,
new HashMap<Integer, Double>());
}
// Add synset link to synterm
tempDictionary.get(synTerm).put(synTermRank,
synsetScore);
}
}
}
// Go through all the terms.
for (Map.Entry<String, HashMap<Integer, Double>> entry : tempDictionary
.entrySet()) {
String word = entry.getKey();
Map<Integer, Double> synSetScoreMap = entry.getValue();
// Calculate weighted average. Weigh the synsets according to
// their rank.
// Score= 1/2*first + 1/3*second + 1/4*third ..... etc.
// Sum = 1/1 + 1/2 + 1/3 ...
double score = 0.0;
double sum = 0.0;
for (Map.Entry<Integer, Double> setScore : synSetScoreMap
.entrySet()) {
score += setScore.getValue() / (double) setScore.getKey();
sum += 1.0 / (double) setScore.getKey();
}
score /= sum;
dictionary.put(word, score);
}
} catch (Exception e) {
e.printStackTrace();
} finally {
if (csv != null) {
csv.close();
}
}
}
public double extract(String word, String pos) {
return dictionary.get(word + "#" + pos);
}
public static void main(String [] args) throws IOException {
if(args.length<1) {
System.err.println("Usage: java SentiWordNetDemoCode <pathToSentiWordNetFile>");
return;
}
String pathToSWN = args[0];
SentiWordNetDemoCode sentiwordnet = new SentiWordNetDemoCode(pathToSWN);
System.out.println("good#a "+sentiwordnet.extract("good", "a"));
System.out.println("bad#a "+sentiwordnet.extract("bad", "a"));
System.out.println("blue#a "+sentiwordnet.extract("blue", "a"));
System.out.println("blue#n "+sentiwordnet.extract("blue", "n"));
}
}
We can pass the pos to sentiwordnet parser.
Download pattern python module
from pattern.en import wordnet
print wordnet.synsets("kill",pos="VB")[0].weight
wordnet.synsets returns list of synsets
and from that we are selecting 1st item
Output will be a tuple of (polarity,subjectivity)
Hope this helps...
I'm using the Stanford Named Entity Recognizer http://nlp.stanford.edu/software/CRF-NER.shtml and it's working fine. This is
List<List<CoreLabel>> out = classifier.classify(text);
for (List<CoreLabel> sentence : out) {
for (CoreLabel word : sentence) {
if (!StringUtils.equals(word.get(AnswerAnnotation.class), "O")) {
namedEntities.add(word.word().trim());
}
}
}
However the problem I'm finding is identifying names and surnames. If the recognizer encounters "Joe Smith", it is returning "Joe" and "Smith" separately. I'd really like it to return "Joe Smith" as one term.
Could this be achieved through the recognizer maybe through a configuration? I didn't find anything in the javadoc till now.
Thanks!
This is because your inner for loop is iterating over individual tokens (words) and adding them separately. You need to change things to add whole names at once.
One way is to replace the inner for loop with a regular for loop with a while loop inside it which takes adjacent non-O things of the same class and adds them as a single entity.*
Another way would be to use the CRFClassifier method call:
List<Triple<String,Integer,Integer>> classifyToCharacterOffsets(String sentences)
which will give you whole entities, which you can extract the String form of by using substring on the original input.
*The models that we distribute use a simple raw IO label scheme, where things are labeled PERSON or LOCATION, and the appropriate thing to do is simply to coalesce adjacent tokens with the same label. Many NER systems use more complex labels such as IOB labels, where codes like B-PERS indicates where a person entity starts. The CRFClassifier class and feature factories support such labels, but they're not used in the models we currently distribute (as of 2012).
The counterpart of the classifyToCharacterOffsets method is that (AFAIK) you can't access the label of the entities.
As proposed by Christopher, here is an example of a loop which assembles "adjacent non-O things". This example also counts the number of occurrences.
public HashMap<String, HashMap<String, Integer>> extractEntities(String text){
HashMap<String, HashMap<String, Integer>> entities =
new HashMap<String, HashMap<String, Integer>>();
for (List<CoreLabel> lcl : classifier.classify(text)) {
Iterator<CoreLabel> iterator = lcl.iterator();
if (!iterator.hasNext())
continue;
CoreLabel cl = iterator.next();
while (iterator.hasNext()) {
String answer =
cl.getString(CoreAnnotations.AnswerAnnotation.class);
if (answer.equals("O")) {
cl = iterator.next();
continue;
}
if (!entities.containsKey(answer))
entities.put(answer, new HashMap<String, Integer>());
String value = cl.getString(CoreAnnotations.ValueAnnotation.class);
while (iterator.hasNext()) {
cl = iterator.next();
if (answer.equals(
cl.getString(CoreAnnotations.AnswerAnnotation.class)))
value = value + " " +
cl.getString(CoreAnnotations.ValueAnnotation.class);
else {
if (!entities.get(answer).containsKey(value))
entities.get(answer).put(value, 0);
entities.get(answer).put(value,
entities.get(answer).get(value) + 1);
break;
}
}
if (!iterator.hasNext())
break;
}
}
return entities;
}
I had the same problem, so I looked it up, too. The method proposed by Christopher Manning is efficient, but the delicate point is to know how to decide which kind of separator is appropriate. One could say only a space should be allowed, e.g. "John Zorn" >> one entity. However, I may find the form "J.Zorn", so I should also allow certain punctuation marks. But what about "Jack, James and Joe" ? I might get 2 entities instead of 3 ("Jack James" and "Joe").
By digging a bit in the Stanford NER classes, I actually found a proper implementation of this idea. They use it to export entities under the form of single String objects. For instance, in the method PlainTextDocumentReaderAndWriter.printAnswersTokenizedInlineXML, we have:
private void printAnswersInlineXML(List<IN> doc, PrintWriter out) {
final String background = flags.backgroundSymbol;
String prevTag = background;
for (Iterator<IN> wordIter = doc.iterator(); wordIter.hasNext();) {
IN wi = wordIter.next();
String tag = StringUtils.getNotNullString(wi.get(AnswerAnnotation.class));
String before = StringUtils.getNotNullString(wi.get(BeforeAnnotation.class));
String current = StringUtils.getNotNullString(wi.get(CoreAnnotations.OriginalTextAnnotation.class));
if (!tag.equals(prevTag)) {
if (!prevTag.equals(background) && !tag.equals(background)) {
out.print("</");
out.print(prevTag);
out.print('>');
out.print(before);
out.print('<');
out.print(tag);
out.print('>');
} else if (!prevTag.equals(background)) {
out.print("</");
out.print(prevTag);
out.print('>');
out.print(before);
} else if (!tag.equals(background)) {
out.print(before);
out.print('<');
out.print(tag);
out.print('>');
}
} else {
out.print(before);
}
out.print(current);
String afterWS = StringUtils.getNotNullString(wi.get(AfterAnnotation.class));
if (!tag.equals(background) && !wordIter.hasNext()) {
out.print("</");
out.print(tag);
out.print('>');
prevTag = background;
} else {
prevTag = tag;
}
out.print(afterWS);
}
}
They iterate over each word, checking if it has the same class (answer) than the previous, as explained before. For this, they take advantage of the fact expressions considered as not being entities are flagged using the so-called backgroundSymbol (class "O"). They also use the property BeforeAnnotation, which represents the string separating the current word from the previous one. This last point allows solving the problem I initially raised, regarding the choice of an appropriate separator.
Code for the above:
<List> result = classifier.classifyToCharacterOffsets(text);
for (Triple<String, Integer, Integer> triple : result)
{
System.out.println(triple.first + " : " + text.substring(triple.second, triple.third));
}
List<List<CoreLabel>> out = classifier.classify(text);
for (List<CoreLabel> sentence : out) {
String s = "";
String prevLabel = null;
for (CoreLabel word : sentence) {
if(prevLabel == null || prevLabel.equals(word.get(CoreAnnotations.AnswerAnnotation.class)) ) {
s = s + " " + word;
prevLabel = word.get(CoreAnnotations.AnswerAnnotation.class);
}
else {
if(!prevLabel.equals("O"))
System.out.println(s.trim() + '/' + prevLabel + ' ');
s = " " + word;
prevLabel = word.get(CoreAnnotations.AnswerAnnotation.class);
}
}
if(!prevLabel.equals("O"))
System.out.println(s + '/' + prevLabel + ' ');
}
I just wrote a small logic and it's working fine. what I did is group words with same label if they are adjacent.
Make use of the classifiers already provided to you. I believe this is what you are looking for:
private static String combineNERSequence(String text) {
String serializedClassifier = "edu/stanford/nlp/models/ner/english.all.3class.distsim.crf.ser.gz";
AbstractSequenceClassifier<CoreLabel> classifier = null;
try {
classifier = CRFClassifier
.getClassifier(serializedClassifier);
} catch (ClassCastException e) {
// TODO Auto-generated catch block
e.printStackTrace();
} catch (ClassNotFoundException e) {
// TODO Auto-generated catch block
e.printStackTrace();
} catch (IOException e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
System.out.println(classifier.classifyWithInlineXML(text));
// FOR TSV FORMAT //
//System.out.print(classifier.classifyToString(text, "tsv", false));
return classifier.classifyWithInlineXML(text);
}
Here is my full code, I use Stanford core NLP and write algorithm to concatenate Multi Term names.
import edu.stanford.nlp.ling.CoreAnnotations;
import edu.stanford.nlp.ling.CoreLabel;
import edu.stanford.nlp.pipeline.Annotation;
import edu.stanford.nlp.pipeline.StanfordCoreNLP;
import edu.stanford.nlp.util.CoreMap;
import org.apache.log4j.Logger;
import java.util.ArrayList;
import java.util.List;
import java.util.Properties;
/**
* Created by Chanuka on 8/28/14 AD.
*/
public class FindNameEntityTypeExecutor {
private static Logger logger = Logger.getLogger(FindNameEntityTypeExecutor.class);
private StanfordCoreNLP pipeline;
public FindNameEntityTypeExecutor() {
logger.info("Initializing Annotator pipeline ...");
Properties props = new Properties();
props.setProperty("annotators", "tokenize, ssplit, pos, lemma, ner");
pipeline = new StanfordCoreNLP(props);
logger.info("Annotator pipeline initialized");
}
List<String> findNameEntityType(String text, String entity) {
logger.info("Finding entity type matches in the " + text + " for entity type, " + entity);
// create an empty Annotation just with the given text
Annotation document = new Annotation(text);
// run all Annotators on this text
pipeline.annotate(document);
List<CoreMap> sentences = document.get(CoreAnnotations.SentencesAnnotation.class);
List<String> matches = new ArrayList<String>();
for (CoreMap sentence : sentences) {
int previousCount = 0;
int count = 0;
// traversing the words in the current sentence
// a CoreLabel is a CoreMap with additional token-specific methods
for (CoreLabel token : sentence.get(CoreAnnotations.TokensAnnotation.class)) {
String word = token.get(CoreAnnotations.TextAnnotation.class);
int previousWordIndex;
if (entity.equals(token.get(CoreAnnotations.NamedEntityTagAnnotation.class))) {
count++;
if (previousCount != 0 && (previousCount + 1) == count) {
previousWordIndex = matches.size() - 1;
String previousWord = matches.get(previousWordIndex);
matches.remove(previousWordIndex);
previousWord = previousWord.concat(" " + word);
matches.add(previousWordIndex, previousWord);
} else {
matches.add(word);
}
previousCount = count;
}
else
{
count=0;
previousCount=0;
}
}
}
return matches;
}
}
Another approach to deal with multi words entities.
This code combines multiple tokens together if they have the same annotation and go in a row.
Restriction:
If the same token has two different annotations, the last one will be saved.
private Document getEntities(String fullText) {
Document entitiesList = new Document();
NERClassifierCombiner nerCombClassifier = loadNERClassifiers();
if (nerCombClassifier != null) {
List<List<CoreLabel>> results = nerCombClassifier.classify(fullText);
for (List<CoreLabel> coreLabels : results) {
String prevLabel = null;
String prevToken = null;
for (CoreLabel coreLabel : coreLabels) {
String word = coreLabel.word();
String annotation = coreLabel.get(CoreAnnotations.AnswerAnnotation.class);
if (!"O".equals(annotation)) {
if (prevLabel == null) {
prevLabel = annotation;
prevToken = word;
} else {
if (prevLabel.equals(annotation)) {
prevToken += " " + word;
} else {
prevLabel = annotation;
prevToken = word;
}
}
} else {
if (prevLabel != null) {
entitiesList.put(prevToken, prevLabel);
prevLabel = null;
}
}
}
}
}
return entitiesList;
}
Imports:
Document: org.bson.Document;
NERClassifierCombiner: edu.stanford.nlp.ie.NERClassifierCombiner;