Parse Whatsapp Log-File in Java

Question

I´m currently working on a little tool which analyses the usage of a group-chat in Whatsapp.

i´m trying to realize it with the whatsapp logfile. I managed it to format the raw .txt to the following format to work with the formated text:

29. Jan. 12:01 - Random Name: message text
29. Jan. 12:22 - Random Name: message text
29. Jan. 12:24 - Random Name: message text
29. Jan. 12:38 - Random Name: message text
29. Jan. 12:52 - Random Name: message text

so far, so good. The Problem is that there are a few floppy lines like:

29. Jan. 08:42 - Random Name2: message text 1
                 additional text of the message 1
29. Jan. 08:43 - Random Name2: message text 2

or even worse:

15. Jan. 14:00 - Random Name: First part of the message
                 second part
                 third part
                 forth part
                 fifth part    
29. Jan. 08:43 - Random Name2: message text 2

I guess I need a kind of algorythm to solve this problem, but i´m pretty new in programming and can´t create such a complex algorithm.

The same problem in Python: parse a whatsApp conversation log

[EDIT]

This is my code which doesn´t work. (I know it´s pretty bad)

import java.io.BufferedReader;
import java.io.BufferedWriter;
import java.io.FileNotFoundException;
import java.io.FileReader;
import java.io.FileWriter;
import java.io.IOException;

public class FormatList {

    public static void main(String[] args) throws IOException {
        // TODO Auto-generated method stub
        FileReader fr = new FileReader("Whatsapp_formated.txt");
        BufferedReader br = new BufferedReader(fr);

        FileWriter fw = new FileWriter("Whatsapp_formated2.txt");
        BufferedWriter ausgabe = new BufferedWriter(fw);

        String line="";
        String buffer="";

        while((line = br.readLine())!=null)
        {
            System.out.println("\n"+line);

            if(line.isEmpty())
            {

            }
            else{
                if(line.charAt(0)=='0'||line.charAt(0)=='1'||line.charAt(0)=='2'||line.charAt(0)=='3'||line.charAt(0)=='4'||line.charAt(0)=='5'||line.charAt(0)=='6'||line.charAt(0)=='7'||line.charAt(0)=='8'||line.charAt(0)=='9')
                {
                    buffer = line;

                }
                else
                {
                    buffer += line;
                }

                 ausgabe.write(buffer);
                 ausgabe.newLine();
                System.out.println(buffer);
            }

            ausgabe.close();

        }




    }

}

[EDIT 2]

In the end i want to read out the file and analyse each line:

29. Jan. 12:01 - Random Name: message text

I can tell when it was sent, who sent it and what/how much he wrote

If i now get the following line:

additional text of the message 1

I neither can tell when it was written nor who sent it

What is the final goal? So what do you want to do? What output do you expect and what did you already write in java to analyze the text? — Nitram, Jun 26 '15 at 14:05
@Eric Pro You can send the log via Whatsapp default settings. — MegaCleptomaniac, Jun 26 '15 at 14:06
@Nitram I want an '.txt' file where all lines are formated properly to read each line and analyse which person wrote how much text. I managed it to reach above listed form with a few "bad" parts like i described — MegaCleptomaniac, Jun 26 '15 at 14:08
You could open the output file with a hex editor and look for "0A" (and "0D 0A"). EDIT: for your code that means, before you are checking if the line is empty use line=line.replace("\r","").replace("\n",""); — Eric, Jun 26 '15 at 14:17
Your input already looks formatted. If you give an idea on how you want your output to look like may be it will help — Johnson Abraham, Jun 26 '15 at 14:26
@EricPro It worked fine for the empty spaces but the problem still is that there are lines without timestamps and names. could i solve this problem in any other way ? — MegaCleptomaniac, Jun 26 '15 at 14:29
@JohnsonAbraham The output should look like the first piece of code given: every line in the formated '.txt' should start with the timestamp followed by the name and the message. No line should remain with just message text bus. I need to somehow either add the timestamp and name to the floppy lines or add the floppy lines to the complete line before. — MegaCleptomaniac, Jun 26 '15 at 14:34

score 2 · Accepted Answer · answered Jun 26 '15 at 16:05

well, I came up with a solution for your problem, I believe, according to what I understood.

Given a file with this format:

29. Jan. 12:01 - Random Name: message text
29. Jan. 12:22 - Random Name: message text
29. Jan. 12:24 - Random Name: message text
29. Jan. 12:38 - Random Name: message text
29. Jan. 12:52 - Random Name: message text
29. Jan. 08:42 - Random Name2: message text 1
                 additional text of the message 1
29. Jan. 08:43 - Random Name2: message text 2
15. Jan. 14:00 - Random Name: First part of the message
                 second part
                 third part
                 forth part
                 fifth part    
29. Jan. 08:43 - Random Name2: message text 2

(This is a file called "wsp.log" in my "data" folder. So the path to access to it is "data/wsp.log")

I expect something like this:

29. Jan. 12:01 - Random Name: message text
29. Jan. 12:22 - Random Name: message text
29. Jan. 12:24 - Random Name: message text
29. Jan. 12:38 - Random Name: message text
29. Jan. 12:52 - Random Name: message text
29. Jan. 08:42 - Random Name2: message text 1 additional text of the message 1
29. Jan. 08:43 - Random Name2: message text 2
15. Jan. 14:00 - Random Name: First part of the message second part third part forth part fifth part
29. Jan. 08:43 - Random Name2: message text 2

According to that, I implemented the following class:

public class LogReader {

    public void processWspLogFile() throws IOException {
        //a. I would reference to my file
        File wspLogFile = new File("data/wsp.log");
        //b. I would use the mechanism to read the file using BufferedReader
        BufferedReader bufferedReader = new BufferedReader(new FileReader(wspLogFile));

        String currLine = null;//This is the current line (like my cursor)

        //This will hold the data of the file in String format
        StringBuilder stringFormatter = new StringBuilder();
        boolean firstIterationDone = false;//The first line will always contains the format, so I will always append it, from the second I will start making the checkings...

        // Now I can use some regex (I'm not really good at this stuff, I just used a Web Page: http://txt2re.com/)
        /* This regex will match the lines that contains the date in this format "29. Jan. 12:22", when I take a look at your file
          I can see that the "additional text of the message" does not contains any date, so I can use that as my point of separation*/
        String regex = "(\\d)(\\d)(\\.)(\\s+)([a-z])([a-z])([a-z])(\\.)(\\s+)(\\d)(\\d)(:)(\\d)(\\d)";
        //As part of using regex, I would like to create a Pattern to make the lines on the list match this expression      
        Pattern wspLogDatePattern = Pattern.compile(regex, Pattern.CASE_INSENSITIVE | Pattern.DOTALL);

        //Use of the line separator of the O.S
        String lineSeparator = System.getProperty("line.separator");

        while ((currLine = bufferedReader.readLine()) != null) {

            if (!firstIterationDone) {
                stringFormatter.append(currLine);
                firstIterationDone = true;
            } else {
                Matcher wspLogDateMatcher = wspLogDatePattern.matcher(currLine);    

                //The first time we will check if the second line has the pattern, if it does, we append a line separator
                if (wspLogDateMatcher.find()) {
                    //It is a "normal" line
                    stringFormatter.append(lineSeparator).append(currLine);             
                } else {
                    //But if it doesn't, we append it on the same line
                    stringFormatter.append(" ").append(currLine.trim());
                }
            }
        }
        System.out.println(stringFormatter.toString());
    }
}

Which I will invoke this way:

public static void main(String[] args) throws IOException {
    new LogReader().processWspLogFile();
}

Hope this can give you some idea or can be useful for your purposes. I know some improvements are needed, refactor is always needed for code :), but by now it can achieve the format expected. Happy coding :).

score 0 · Answer 2 · answered Jun 26 '15 at 14:59

Try this code. And see if it does whats intended.

import java.io.BufferedReader;
import java.io.BufferedWriter;
import java.io.FileReader;
import java.io.FileWriter;
import java.io.IOException;

public class WhatsappFormatted {

    public static void main(String[] args) throws IOException {
        // TODO Auto-generated method stub
        char preString = '-';
        char searchString = ':';
        FileReader fr = new FileReader("Whatsapp_formated.txt");
        BufferedReader br = new BufferedReader(fr);

        FileWriter fw = new FileWriter("Whatsapp_formated2.txt");
        BufferedWriter ausgabe = new BufferedWriter(fw);

        String line = "";
        String buffer = "";
        String lastMember = null;
        while ((line = br.readLine()) != null) {
            System.out.println("\n" + line);

            if (!line.isEmpty())

                if (Character.isDigit(line.charAt(0)) && Character.isDigit(line.charAt(1))) {
                    lastMember = line.substring(0, line.indexOf(searchString, line.indexOf(preString)) + 1);
                    buffer = line.trim();
                } else {
                    buffer += "\n" + lastMember + line.trim();
                }

            ausgabe.write(buffer);
            ausgabe.newLine();
            System.out.println(buffer);
        }

        ausgabe.close();

    }

}

with this code the size of the new document is like 20 times bigger than my old file due to a String which gets bigger with every line in the unformated version. — MegaCleptomaniac, Jun 29 '15 at 06:23
That will be the case with formatted and non-formatted output since in the formatted output we will be putting in our stuff, in this case we will be adding this '15. Jan. 14:00 - Random Name:' to each line. But is this code serving your purpose. — Johnson Abraham, Jun 29 '15 at 07:18

Parse Whatsapp Log-File in Java

2 Answers2