Writing a script to extract pictures from Word documents

· August 15, 2017

I had a problem and I noticed that I’ve, in the last couple of years, started to think differently about how to solve problems like these. I thought I share the solution to my problem here but also a little bit about the reasoning behind my problem solving.

The problem is easy enough to describe: I wanted to extract all the images from 20+ Word documents. I decided to write a script and share it here.

TL;DR - just the script

Here’s the bash script that I ended up with

#!/bin/bash
rm -rf zips
mkdir zips
cp docs/*.docx zips

for file in ./zips/*.docx; do
  mv "$file" $file.zip
  unzip $file.zip 'word/media/*.jpeg' -d $file.images
  rm $file.zip
done

Decisions, decisions, decisions

First of all I decided to write a scrip to do this. I did that because I want to keep my programming thinking alive and also learn something a little new (almost) every day. I suck at shell scripting and this seemed like a nice exercise.

The first thing I did when writing the script was to think about how I could run the script over and over again, in an easy fashion. For me that means that I kept the raw data (the Word documents in this case) in one folder and the output in another.

I’ve learned, the hard way, that writing any kind of code requires ~a few~ many iterations. I want to test an idea, run that part, tweak the idea and then run again. Many times a minute.

For that reason I set myself up in away that support this iteration. For that reason I clean up the output folder in the first set, like this: rm -rf zips.

The actual algorithm builds on the fact that Word documents (.docx format) are actually zip files. My first step is therefor to copy them to my working directory zips with the command cp docs/*.docx zips

Once the documents are in the working directory I need to change the extension of each file, from .docx to .zip. I found no easy way to do this with a single command, but rather ended up iterating over them.

This is accomplished with the for construct, and in all honesty I’ve never tried that before in shell scripts. The block inside do to done gets iterated for each file in the directory.

Renaming each file turns out to be easy inside the loop, by simply appending .zip to the end of the current file name mv "$file" $file.zip

Now that I’m already iterating I decided to unzip each directory too. For this I’m using the unzip command. The word/media/*.jpeg part means that I only unzip that folder inside the zip-archive. -d $file.images creates a folder with the suffix .images where the images are extracted too.

Finally I decided to clean up and remove the zip archive, inside the loop. That is a little bit unnecessary but why not.

Summary

This was a fun little exercise.

  • It was faster to code this up than to open and extract all images manually
  • I can do this over and over. I might add more pictures to the documents later, or more documents…
  • I learned some new stuff
  • I got to blog about it too.

Twitter, Facebook