标注的Trigger的位置是否与句子中的位置相同

最新推荐文章于 2024-09-25 10:59:43 发布

小妖精Fsky

最新推荐文章于 2024-09-25 10:59:43 发布

阅读量562

点赞数

分类专栏： Text Process 文章标签： java 语料处理

本文链接：https://blog.csdn.net/appleml/article/details/50475479

版权

Text Process 专栏收录该内容

6 篇文章 1 订阅

订阅专栏

package AboutProtein;

import java.io.BufferedReader;
import java.io.BufferedWriter;
import java.io.File;
import java.io.FileReader;
import java.io.FileWriter;
import java.io.IOException;
import java.util.ArrayList;
import java.util.List;

public class LocRight
{
	public static void main(String args[])
	{
		try
		{			
			File file = new File("F:\\generate jar\\2013_BioEvent");
			File[] files = file.listFiles();
			for(int fileId = 0; fileId < files.length; fileId++)
			{
				String fileName = files[fileId].getName();
				FileReader frAll = new FileReader(files[fileId]);
				BufferedReader brAll = new BufferedReader(frAll);
				
				FileWriter fw = new FileWriter("F:\\generate jar\\2013BioEvent\\" + fileName);
				BufferedWriter bw = new BufferedWriter(fw);
				
				List<Protein> proAll = new ArrayList<Protein>();
				List<Protein> proPart = new ArrayList<Protein>();
				List<Trigger> triList = new ArrayList<Trigger>();
				//List<Trigger> problem = new ArrayList<Trigger>();
				List<String> event = new ArrayList<String>();
				String[] sent = new String[1000];
				List<String> sentList = new ArrayList<String>();
				if(fileName.equals("PMC-1134658-00-TIAB.txt"))
				{
					System.out.println(fileName);
				}
				String str;
				while((str = brAll.readLine()) != null)
				{
					if(str.length() == 0)
					{
						for(int tId = 0; tId < triList.size(); tId++)
						{
							Trigger tr = triList.get(tId);
							int start = tr.triggerStart;
							int end = tr.triggerEnd;
							String name = tr.triggerName;
							if(start == end)
							{
								String word = sentList.get(start);
								if(!word.equals(name))
								{
									System.out.println(fileName);
									System.out.println(name + " " + word);
								}
							}else if(start < end)
							{
								StringBuffer sb = new StringBuffer();
								for(int j = start; j <= end; j++)
								{
									if(j == end)
									{
										sb.append(sentList.get(j));
									}else
									{
										sb.append(sentList.get(j) + " ");
									}
								}
								String word = sb.toString();
								if(!word.equals(name))
								{
									System.out.println(fileName);
									System.out.println(name + " " + word);
								}
							}else 
							{
								System.out.println("have a problem");
							}
						}
						proAll.clear();
						proPart.clear();
						triList.clear();
						sentList.clear();
					}else if(str.startsWith("#"))
					{
						String[] proStr = str.split(" ");
						
						StringBuffer proB = new StringBuffer();				    	
						for(int prosId = 4; prosId < proStr.length; prosId++)
						{
							if(prosId == proStr.length - 1)
							{
								proB.append(proStr[prosId]);
							}else
							{
								proB.append(proStr[prosId] + " ");
							}
						}	
						String proName = proB.toString();						
						Protein prot = new Protein();
						prot.proId = proStr[1];
						prot.proStart = Integer.parseInt(proStr[2]);
						prot.proEnd = Integer.parseInt(proStr[3]);
						prot.proName = proName;						
						proAll.add(prot);
						
					}else if(str.startsWith("$"))
					{
                        String[] proStr = str.split(" ");
						
						StringBuffer proB = new StringBuffer();				    	
						for(int prosId = 4; prosId < proStr.length; prosId++)
						{
							if(prosId == proStr.length - 1)
							{
								proB.append(proStr[prosId]);
							}else
							{
								proB.append(proStr[prosId] + " ");
							}
						}	
						String proName = proB.toString();						
						Protein prot = new Protein();
						prot.proId = proStr[1];
						prot.proStart = Integer.parseInt(proStr[2]);
						prot.proEnd = Integer.parseInt(proStr[3]);
						prot.proName = proName;						
						proPart.add(prot);
					}else if(str.startsWith("@"))
					{
						String[] trit = str.split(" ");
						
						StringBuffer triB = new StringBuffer();				    	
						for(int prosId = 5; prosId < trit.length; prosId++)
						{
							if(prosId == trit.length - 1)
							{
								triB.append(trit[prosId]);
							}else
							{
								triB.append(trit[prosId] + " ");
							}
						}	
						String triName = triB.toString();
						
						Trigger trig = new Trigger();
						trig.triggerId = trit[1];
						trig.triggerType = trit[2];
						trig.triggerStart = Integer.parseInt(trit[3]);
						trig.triggerEnd = Integer.parseInt(trit[4]);
						trig.triggerName = triName;
						triList.add(trig);
						
					}else if(str.startsWith("%"))
					{
						event.add(str);
					}else
					{						
						sent = str.split(" ");
						for(int j = 0; j < sent.length; j++)
						{
							sentList.add(sent[j]);
						}						
					}
				}
				brAll.close();
				frAll.close();
			}
		}catch(IOException io)
		{
			io.printStackTrace();
		}
	}		
}