Document Tools

OCR (Optical Character Recognition)

OCR an image to text:

tesseract image.png output_text

OCR with specific language:

tesseract image.png output_text -l eng

OCR a PDF (convert pages to images first, then OCR):

# Convert PDF pages to images
convert -density 300 input.pdf -quality 100 page-%03d.png
# OCR each page
for f in page-*.png; do tesseract "$f" "${f%.png}" -l eng; done
# Combine results
cat page-*.txt > full_text.txt

Document Conversion

Convert between formats (Word, ePub, HTML, Markdown, etc.):

pandoc input.docx -t markdown -o output.md
pandoc input.epub -t markdown -o output.md
pandoc input.html -t markdown -o output.md
pandoc input.md -t html -o output.html

ePub to plain text:

pandoc input.epub -t plain -o output.txt

Image Processing

Resize an image:

convert input.png -resize 800x800 output.png

Crop an image:

convert input.png -crop WIDTHxHEIGHT+X+Y output.png

Convert image format:

convert input.png output.jpg

Get image metadata:

exiftool image.png

Important Notes

• These tools are installed inside the container. If the container restarts, run: bash /workspace/scripts/install-tools.sh as root to reinstall them.
• For large PDFs, process page by page to avoid memory issues.
• ImageMagick policy may block PDF conversion -- if so, the policy file is at /etc/ImageMagick-6/policy.xml and needs the PDF line commented out.