webservices/compbio/nosql/cassandra/DataParsing.java

   1 package combio.cassandra;
   2
   3 import java.io.BufferedReader;
   4 import java.io.IOException;
   5 import java.io.InputStreamReader;
   6 import java.net.HttpURLConnection;
   7 import java.net.MalformedURLException;
   8 import java.net.URL;
   9 import java.net.URLConnection;
  10 import java.text.ParseException;
  11 import java.text.SimpleDateFormat;
  12 import java.util.Calendar;
  13 import java.util.Date;
  14
  15 public class DataParsing {
  16         private CassandraCreate cc = new CassandraCreate();
  17
  18         public void ParsingTest(String input) {
  19                 System.out.println("Inserting.....");
  20                 URL url, urltable;
  21 //              int count = 0;
  22 //              int countNotM = 0;
  23 //              int countEr = 0;
  24 //              int countEmp = 0;
  25                 Calendar cal = Calendar.getInstance();
  26                 String date = cal.get(Calendar.YEAR) + "/" + cal.get(Calendar.MONTH) + "/" + cal.get(Calendar.DAY_OF_MONTH);
  27                 try {
  28                         url = new URL(input);
  29                         URLConnection conn = url.openConnection();
  30                         BufferedReader br = new BufferedReader(new InputStreamReader(conn.getInputStream()));
  31                         String line;
  32
  33                         while ((line = br.readLine()) != null) {
  34                                 if (line.matches(date + "(.*)jp_[^\\s]+")) {
  35                                         String[] table;
  36                                         table = line.split("\\s+");
  37         //                              count++;
  38                                                 if (!cc.CheckIP(table[table.length - 1])) {
  39         //                                              countNotM++;
  40                                                         urltable = new URL("http://www.compbio.dundee.ac.uk/www-jpred/results/" +table[table.length - 1]+"/"+table[table.length - 1]+".concise.fasta");
  41                                                         HttpURLConnection httpConnection = (HttpURLConnection) urltable.openConnection();
  42                                                         if (httpConnection.getResponseCode() > 200) {
  43 //                                                              countEr++;
  44                                                                 continue;
  45                                                         }
  46                                                         try {
  47                                                                 BufferedReader br1 = new BufferedReader(new InputStreamReader(httpConnection.getInputStream()));
  48                                                                 String lineNext;
  49                                                                 String newProt = "";
  50                                                                 String jnetpred = "";
  51                                                                 while ((lineNext = br1.readLine()) != null) {
  52                                                                         if (lineNext.equals(">QUERY")) {
  53                                                                                 while (!(lineNext = br1.readLine()).matches(">[^\\s]+"))
  54                                                                                         newProt += lineNext;
  55                                                                         } else if (lineNext.equals(">jnetpred")) {
  56                                                                                 while (!(lineNext = br1.readLine()).matches(">[^\\s]+"))
  57                                                                                         jnetpred += lineNext;
  58                                                                         }
  59                                                                 }
  60                                                                 br1.close();
  61                                                                 if (newProt.length() <= 1) {
  62                 //                                                      countEmp++;
  63                                                                         continue;
  64                                                                 }
  65
  66                                                                 SimpleDateFormat formatter = new SimpleDateFormat("yyyy/MM/dd");
  67
  68                                                                 String dateInString1 = table[0].substring(0, table[0].indexOf(":"));
  69
  70                                                                 long dateWork1 = 0;
  71
  72                                                                 try {
  73                                                                         Date dat1 = formatter.parse(dateInString1);
  74                                                                         dateWork1 = dat1.getTime();
  75
  76                                                                 } catch (ParseException e) {
  77                                                                         e.printStackTrace();
  78                                                                 }
  79                                                                 cc.InsertData(dateWork1, table[0], table[1], table[2], table[table.length - 1], "OK", "OK", newProt, jnetpred);
  80                                                         } catch (IOException e) {
  81                                                                 //                      e.printStackTrace();
  82                                                         }
  83                                         //      }
  84                                         }
  85                                 }
  86                         }
  87                         br.close();
  88 //                      System.out.println("Match " + count);
  89 //                      System.out.println("Not Match " + countNotM);
  90 //                      System.out.println("Error " + countEr);
  91 //                      System.out.println("No protein " + countEmp);
  92                 } catch (MalformedURLException e) {
  93                         e.printStackTrace();
  94                 } catch (IOException e) {
  95                         e.printStackTrace();
  96                 }
  97         }
  98 }