The thing to remember is that MS Office documents are just a collection of XML files zipped up. The Java APIs present a virtual filesystem for the file. See https://docs.oracle.com/javase/7/docs/api/index.html?java/util/zip/package-summary.html

Treat the zip contents like you would a regular file on disk -- don't overcomplicate it. Unless the files are very large, you do not need to write anything to disk in your Pega routines -- everything is in memory.

Step 1: Prepare for the Pega coding by unzipping the docx file. Use a command line utility, or else rename the .docx extension to .zip to more easily use your favorite unzip tool. Whatever works. The Java jar command works well...

> jar -tf test.docx

> jar -xvf test.docx

Step 2: Find your the XML file that contains whatever you are interested in. It's just text, so a search tool or IDE will work well. In my simple example the phrase I typed in Word was at word/document.xml.

Step 3: Try this in an IDE like Eclipse or Netbeans. Please be kind to yourself: Unless you are an old pro, don't develop Java inside Pega. Death by a thousand cuts. After you run perform(), exercise zipUp() for fun and create a zip file from thin air.

package replaceme;

import java.io.*;

import java.util.zip.*;

public class TestClass {

    public static void main(String[] args) {

        TestClass tc = new TestClass();

        tc.perform();

        // tc.zipUp();

    }

    private void perform() {

        try {

            ZipInputStream zis = new ZipInputStream(new FileInputStream(new File("path-to-my-file/test.docx")));

            ZipEntry entry = null;

            byte[] b = null;

            while (null != (entry = zis.getNextEntry())) {

                //     System.out.println(entry.getName() + " " + entry.isDirectory());

                if ("word/document.xml".equals(entry.getName())) {

                    int size = (int)entry.getSize();

                    b = new byte[size];

                    zis.read(b, 0, size);

                }

            }

            zis.close();

            String s = new String(b);

            System.out.println(s);

            System.out.println(s.contains("Some text I’m looking for"));

        } catch (Exception e) {

           e.printStackTrace();

        }

    }

    private void zipUp() {

        try {

            ZipOutputStream zos = new ZipOutputStream(new FileOutputStream(new File("path-to-create-my-outputfile/testout.zip")));

            ZipEntry ze1 = new ZipEntry("foo.txt");

            ZipEntry ze2 = new ZipEntry("bar.txt");

            byte[] b = "This little piggy went to market.".getBytes();

            zos.putNextEntry(ze1);

            zos.write(b, 0, b.length);

            b = "This little piggy stayed home.".getBytes();

            zos.putNextEntry(ze2);

            zos.write(b, 0, b.length);

            zos.close();

        } catch(Exception e) {

            e.printStackTrace();

        }

    }

}

Step 4: Once you understand the above, revise this code to work into your Pega activity. You are responsible for obtaining a page reference to your Word document. For this example I attached a file to a work object and examined the clipboard to find the pzInsKey. I won't belabor how Pega stores binary content... it's all in the developer help.

Pay attention to the file encoding of the Base64 decode call. This could vary depending on your desktop settings. If in doubt, UTF-8 might be forgiving. Also note that Pega has a couple functions for UTF-8 decoding, but they wrap only the purely string-based decode calls and are useless here. Consider writing your own.

If you haven't done this before, notice the pageContent variable. This Local Variable is declared on the Parameters tab. Then you can use it outright in a Java step and refer to it in other activity steps with the "local." prefix...

... continued image....

If you run that activity, the last step will look odd because Pega won't understand how to render that XML. Nevertheless, I was able to see the text from my Word document. As for parsing your Word text and mapping the result to clipboard properties, I recommend regular expressions for the former, and using a local variable in the Java step for the latter.

Illustrating all this in an Activity is easiest to understand, but personally I would put the Java in one or two Pega Functions. Maybe work in safety sanity checks for the size of the file or content.

Cheers!