Document-Image-Extractor

CLI tool to extract embedded images from DOCX, PDF, PPTX and XLSX files, with deduplication , size filtering, and batch export to ZIPs.

---

Features:

Extract images from:

• DOCX (Word documents)
• PDF (documents)
• PPTX (Powerpoint documents)
• XLSX (Excel documents)

Outputs:

• Creates a ZIP per input file with extracted images

built-in helpers:

• Deduplication (skips repeated images within the same document)
• Size filter (min_kb default is 5kb)
• Handles “no images” and corrupt files gracefully

---

Project status

this repository is begin improved phase by phase

---

Requirements

• python 3.12+ (recomended)

Dependencies (install from 'requirements.txt'):

• python-docx
• PyMuPDF
• pillow

Installation

1. Clone the repository

git clone https://github.com/LeoMurilloDev/document-image-extractor.git
cd document-image-extractor 

2. Create and activate a virtual environment

Windows

python -m  venv .venv
.\.venv\Scripts\activate

macOS / Linux

python3 -m venv .venv
source .venv/bin/activate

3. Install dependencies

pip install -r requirements.txt

Usage

Folder structure expected by the script

the script creates these folders automatically if they don't exist:

• Entrdas_archivos/ -> place your .docx and .pdf files here
• Salidas_archivos/ -> output ZIPs will be generated here
• temp/ -> temporary extraction folder (auto-cleaned)

Configuration

You can customize filters without editing the code using config.json (repo root). Example:

{
  "filters": {
    "min_kb": 5,
    "min_width": 0,
    "min_height": 0
  }
}

• min_kb: minimum file size in kb (default: 5)
• min_width/ min_height: optional dimension filter (0 disables it)

Run

python main.py

CLI usage

The tool can be used with default folders/config:

python main.py

python main.py --input Entradas_archivos --output Salidas_archivos

python main.py --input example.pptx --output Salidas_archivos

python main.py --input Entradas_archivos --recursive

python main.py --input Entradas_archivos --min-kb 1 --min-width 100 --min-height 100

python main.py --input Entradas_archivos --no-dedup

python main.py --input Entradas_archivos --format folder

python main.py --input Entradas_archivos --log-level DEBUG --log-file logs/debug.log

Output

• For each input file, a ZIP is created in Salidas_archivos/
• Example:
  • Input: Entradas_archivos/report.pdf
  • Output: Salidas_archivos/report.zip

What to expect

When you run the script, it prints a summary per file:

• guardadas -> images saved successfully
• duplicadas -> images skipped due to hash duplication
• pequeñas -> images filtered out by size
• encontradas -> images found inside the document

Important notes

• In DOCX, images are saved using the real extension (.jpg, .png, .gif, etc)
• temp/ is cleaned even when a file fails

Test suites

we use small test suites to validate.

Documents to try

Includes:

• Mixed formats (JPG/PNG/GIF)
• Duplicates
• Small icon filtered out by size
• Corrupt files (error handling) Manual validation steps:

1. Copy test files into Entradas_archivos/
2. Run python main.py
3. Verify
   • Output ZIPs exist in Salidas_archivos/
   • Extencions are correct in DOCX resutls (.jpg, .png, .gif)
   • Duplicates are removed
   • temp/ is empty at the end

Contributing

if you want to propose changes:

1. Fork the repo
2. Create a branch
3. Open a PR with a clear description