我的目标是执行情绪分析，主题提取，实体...当我用thrift服务器启动执行以下java代码时....
package com.diskoverorta.tamanager;
import com.diskoverorta.entities.EntityManager;
//import com.diskoverorta.lifesciences.LSInterface;
import com.diskoverorta.osdep.StanfordNLP;
import com.diskoverorta.utils.EntityUtils;
import com.diskoverorta.vo.*;
import com.google.common.base.Charsets;
import com.google.common.io.Files;
import com.google.gson.Gson;
import com.google.gson.GsonBuilder;
import com.diskoverorta.pyinterface.*;
//import com.serendio.diskoverer.lifesciences.document.LifeScienceDocument;
import java.io.File;
import java.util.List;
import java.util.Set;
import java.util.TreeSet;
import org.kohsuke.args4j.Argument;
import org.kohsuke.args4j.CmdLineException;
import org.kohsuke.args4j.CmdLineParser;
import org.kohsuke.args4j.Option;
import java.io.IOException;
import java.util.ArrayList;
import java.nio.charset.Charset;
public class TextManager
{
@Option(name = "-analyze", aliases = {"-analyze"}, usage = "Accepted options for analyze : Entity, Sentiment, Topic, Keyword, All")
String analysis;
@Option(name = "-Text", aliases = {"-text"}, usage = "Extracts Text analytics components for given Text")
String text;
@Option(name = "-File", aliases = {"-file"}, usage = "Extracts Text analytics components for given File")
String fileName;
@Argument
List<String> arguments = new ArrayList<String>();
static StanfordNLP nlpStanford = null;
static ThriftClient pyClient = null;
DocumentObject doc = null;
public TextManager()
{
if(nlpStanford==null)
{
nlpStanford = new StanfordNLP();
}
if(pyClient==null)
{
pyClient = new ThriftClient("localhost",8002);
}
}
public DocSentObject tagTextAnalyticsComponents(String sDoc,TAConfig config)
{
DocSentObject doc = new DocSentObject();
List<String> sentList = nlpStanford.splitSentencesINDocument(sDoc);
for(int i=0; i < sentList.size(); i++)
{
SentenceObject temp = new SentenceObject();
temp.sentenceText = sentList.get(i);
if(config.analysisConfig.get("Entity") == "TRUE")
temp.entities = (new EntityManager().getSelectedEntitiesForSentence(temp.sentenceText,config.entityConfig));
doc.docSentences.add(temp);
}
return doc;
}
public DocumentObject aggregateDocumentComponentsFromSentences(DocSentObject docSent)
{
DocumentObject docObject = new DocumentObject();
for (SentenceObject sentObj : docSent.docSentences)
{
docObject.entities.currency.addAll(sentObj.entities.currency);
docObject.entities.date.addAll(sentObj.entities.date);
docObject.entities.location.addAll(sentObj.entities.location);
docObject.entities.organization.addAll(sentObj.entities.organization);
docObject.entities.percent.addAll(sentObj.entities.percent);
docObject.entities.person.addAll(sentObj.entities.person);
docObject.entities.time.addAll(sentObj.entities.time);
}
docObject.entitiesMeta = EntityUtils.extractEntityMap(docObject.entities);
return docObject;
}
public String tagTextAnalyticsComponentsINJSON(String sDoc,TAConfig config)
{
DocSentObject doc = tagTextAnalyticsComponents(sDoc,config);
DocumentObject docObject = aggregateDocumentComponentsFromSentences(doc);
Gson gson = new GsonBuilder().setPrettyPrinting().create();
String jsonOutput = gson.toJson(docObject);
return jsonOutput;
}
public String tagUniqueTextAnalyticsComponentsINJSON(String sDoc,TAConfig config)
{
String jsonOutput = "";
APIOutput apiOut = new APIOutput();
Gson gson = new GsonBuilder().setPrettyPrinting().create();
if(config.analysisConfig.get("Entity") == "TRUE")
{
DocSentObject doc = tagTextAnalyticsComponents(sDoc, config);
DocumentObject docObject = aggregateDocumentComponentsFromSentences(doc);
EntityAPISet apiSet = EntityUtils.getEntitySet(docObject);
apiOut.entity_general = apiSet;
}
// if(config.analysisConfig.get("LSEntity") == "TRUE")
// apiOut.entity_lifesciences = gson.fromJson(LSInterface.getLSEntitiesinJSON(sDoc),LifeScienceDocument.class);
if(config.analysisConfig.get("Topic") == "TRUE")
{
Set<String> topic_set = new TreeSet<String>();
if (apiOut.text_information == null)
apiOut.text_information = new TextInformation();
if(pyClient != null)
{
Charset.forName("UTF-8").encode(sDoc);
List<String> topics = pyClient.getTopics(sDoc);
for (String topic : topics)
topic_set.add(topic);
}
if (topic_set.isEmpty() == true)
{
topic_set.add("Topic analyzer not working, Start Thrift server at port 8002");
}
apiOut.text_information.topics = topic_set;
}
if(config.analysisConfig.get("Keyword") == "TRUE")
{
Set<String> keyword_set = new TreeSet<String>();
if (apiOut.text_information == null)
apiOut.text_information = new TextInformation();
if(pyClient != null)
{
List<String> keywords = pyClient.getKeywords(sDoc);
for (String keyword : keywords)
keyword_set.add(keyword);
}
if (keyword_set.isEmpty() == true)
{
keyword_set.add("Keyword extractor not working, Start Thrift server at port 8002");
}
apiOut.text_information.keywords = keyword_set;
}
if(config.analysisConfig.get("Sentiment") == "TRUE")
{
String senti_temp = null;
Set<String> senti_set = new TreeSet<String>();
if (apiOut.text_information == null)
apiOut.text_information = new TextInformation();
if(pyClient != null)
{
if (apiOut.text_information == null)
apiOut.text_information = new TextInformation();
if (config.sentimentConfig.get("textType") == "blogs_news")
senti_temp = pyClient.getSentimentScore(sDoc, config.sentimentConfig.get("title"), config.sentimentConfig.get("middleParas"), config.sentimentConfig.get("lastPara"), 1);
if (config.sentimentConfig.get("textType") == "reviews")
senti_temp = pyClient.getSentimentScore(sDoc, config.sentimentConfig.get("title"), config.sentimentConfig.get("topDomain"), config.sentimentConfig.get("subDomain"));
else
senti_temp = pyClient.getSentimentScore(sDoc, config.sentimentConfig.get("textType"));
}
if (senti_temp == null)
{
senti_temp = "Sentiment analyzer not working, Start Thrift server at port 8002";
}
senti_set.add(senti_temp);
apiOut.text_information.sentiment = senti_set;
}
return gson.toJson(apiOut);
}
public static void main(String args[])
{
TAConfig config = new TAConfig();
TextManager temp = new TextManager();
CmdLineParser parser = new CmdLineParser(temp);
String sample = "The last week of Parliament is unlikely to see any business being conducted with the Congress party showing no signs of relenting on its demands and continuing to seek the resignation of Union and state ministers, while hitting out at the BJP for its “politics of abuse.Congress leaders parried questions on whether the MPs would continue to show placards in the House, an act that had angered the presiding officer. But Monday morning saw several Congress MPs carry placards to the House.Here’s the latest from Parliament:3:22 pm: Lok Sabha adjourned for the day amid continued Opposition protests over Lalitgate and Vyapam issue.2:38 pm: Rajya Sabha adjourned for the day after uproar over Lalit Modi, Vyapam and some other issues.: http://indianexpress.com/article/india/politics/parliament-logjam-25-suspended-mps-return-will-congress-let-house-function/#sthash.63UHi4eu.dpuf";
String trialtext = "";
try {
parser.parseArgument(args);
}catch(CmdLineException ex)
{
ex.printStackTrace();
}
System.out.println("Command line API :");
parser.printUsage(System.out);
config.entityConfig.put("person", "TRUE");
config.entityConfig.put("organization", "TRUE");
config.entityConfig.put("location", "TRUE");
config.entityConfig.put("date", "TRUE");
config.entityConfig.put("time", "TRUE");
config.entityConfig.put("currency", "TRUE");
config.entityConfig.put("percent", "TRUE");
if((temp.text == null) && (temp.fileName == null)) {
trialtext = "The last week of Parliament is unlikely to see any business being conducted with the Congress party showing no signs of relenting on its demands and continuing to seek the resignation of Union and state ministers, while hitting out at the BJP for its “politics of abuse.Congress leaders parried questions on whether the MPs would continue to show placards in the House, an act that had angered the presiding officer. But Monday morning saw several Congress MPs carry placards to the House.Here’s the latest from Parliament:3:22 pm: Lok Sabha adjourned for the day amid continued Opposition protests over Lalitgate and Vyapam issue.2:38 pm: Rajya Sabha adjourned for the day after uproar over Lalit Modi, Vyapam and some other issues.: http://indianexpress.com/article/india/politics/parliament-logjam-25-suspended-mps-return-will-congress-let-house-function/#sthash.63UHi4eu.dpuf";
System.out.println("No Text source provided considering this default text for analysis : "+trialtext);
}
else if((temp.text != null) && (temp.fileName != null)) {
System.out.println("Command line text input considered for Text Analytics : " + temp.text);
trialtext = temp.text;
}
else if((temp.text != null)) {
System.out.println("Command line text input considered for Text Analytics : " + temp.text);
trialtext = temp.text;
}
else if((temp.fileName != null)) {
System.out.println("Command line text input considered for Text Analytics");
try {
trialtext = Files.toString(new File(temp.fileName), Charsets.UTF_8);
}catch(IOException ex)
{
ex.printStackTrace();
}
System.out.println("File input considered for Text Analytics : " + trialtext);
}
if((temp.analysis != null) && (temp.analysis.equalsIgnoreCase("Entity") == true)) {
config.analysisConfig.put("Entity", "TRUE");
System.out.println("Analyzing Entity in given text");
}
else if ((temp.analysis != null) && (temp.analysis.equalsIgnoreCase("Sentiment") == true)) {
config.analysisConfig.put("Sentiment", "TRUE");
System.out.println("Analyzing Sentiment in given text");
}
else if ((temp.analysis != null) && (temp.analysis.equalsIgnoreCase("Topic") == true)) {
config.analysisConfig.put("Topic", "TRUE");
System.out.println("Analyzing Topic in given text");
}
else if ((temp.analysis != null) && (temp.analysis.equalsIgnoreCase("Keyword") == true)) {
config.analysisConfig.put("Keyword", "TRUE");
System.out.println("Extracting keywords in given text");
}
else if ((temp.analysis != null) && (temp.analysis.equalsIgnoreCase("All") == true)) {
System.out.println("Analyzing Entity, Sentiment, Topic and Keywords");
config.analysisConfig.put("Entity","TRUE");
config.analysisConfig.put("Sentiment", "TRUE");
config.analysisConfig.put("Topic", "TRUE");
config.analysisConfig.put("Keyword","TRUE");
}
else {
System.out.println("Analysis options not specified. Possible values : Entity, Sentiment, Topic, Kaeyword, All");
System.out.println("Choosing All by default");
config.analysisConfig.put("Entity","TRUE");
config.analysisConfig.put("Sentiment", "TRUE");
config.analysisConfig.put("Topic", "TRUE");
config.analysisConfig.put("Keyword","TRUE");
}
System.out.println("Text Analytics output : ");
System.out.println(temp.tagUniqueTextAnalyticsComponentsINJSON(trialtext, config));
}
}
我得到这样的错误......
位于org.apache.thrift.transport.TTransport.readAll（TTransport.java:86）的org.apache.thrift.transport.TIOStreamTransport.read（TIOStreamTransport.java:132）上的org.apache.thrift.transport.TTransportException在org.apache.thrift.protocol.TBinaryProtocol.readAll（TBinaryProtocol.java:429）在org.apache.thrift.protocol.TBinaryProtocol.readI32（TBinaryProtocol.java:318）在org.apache.thrift.protocol.TBinaryProtocol.readMessageBegin （TBinaryProtocol.java:219）位于com.diskoverorta.pyinterface的com.diskoverorta.pyinterface.PyInterface $ Client.recv_getTopics（PyInterface.java:142）的org.apache.thrift.TServiceClient.receiveBase（TServiceClient.java:69）。 PyInterface $ Client.getTopics（PyInterface.java:129）在com.diskoverorta.pyinterface.ThriftClient.getTopics（ThriftClient.java:123）在com.diskoverorta.tamanager.TextManager.tagUniqueTextAnalyticsComponentsINJSON（TextManager.java:142）在com.diskoverorta .tamanager.TextManager.main（TextManager.java:285）线程“main”中的异常java.lang.NullPointerExcep在com.diskoverorta.tamanager.TextManager.tagUniqueTextAnalyticsComponentsINJSON（TextManager.java:145）的com.diskoverorta.tamanager.TextManager.main（TextManager.java:285）
我对此完全陌生....我无法弄清楚问题所在......请帮助，谢谢!!
答案 0 :(得分：0)
似乎config.analysisConfig为空。因此在调用config.analysisConfig.get
你应该添加所有使用config.analysisConfig的ifs：
config.analysisConfig != null && as：
if(config.analysisConfig != null && config.analysisConfig.get("Entity") == "TRUE") {
...
}