OCR PDF document using Tesseract and compress using JBIG2 lossy compression

pdf-ocr-compress.sh

#!/bin/bash
SCRIPT_NAME="OCR and jbig2 compress document using Tesseract"
VERSION_NUMBER="v2024-02-02"
AUTHOR="Martin Brozkeff Malec - https://brozkeff.net"
LICENSE="MIT (https://opensource.org/license/mit/)"
# Shared on https://gist.github.com/brozkeff/76a8c537b15458a58355949c4c70da08

echo "${SCRIPT_NAME} ${VERSION_NUMBER}"
echo "Author: ${AUTHOR}"
echo "License: ${LICENSE}"

# This script performs OCR on a PDF file with optional compression to jbig2 lossy and saves the result
# with a '-OCR' suffix. It supports multiple languages, with packages for Czech, English, and Tibetan required.
# Before running, ensure the necessary Tesseract OCR language packages are installed.

# Language packages needed: tesseract-ocr-ces (Czech), tesseract-ocr-eng (English), tesseract-ocr-script-tibt (Tibetan),
# and the main package tesseract-ocr. The jbig2enc package is also required for jbig2 lossy compression.

# Languages to process: To add or remove languages, modify the LANGS variable and REQUIRED_PKGS for Tesseract packages.
# Optimization level: default is 1. Options are 0-3. For lossy compression, 2-3 is required.
# REDO_OCR is 1 to force making new OCR.
LANGS="ces+eng+Tibetan"
REQUIRED_PKGS=("tesseract-ocr" "tesseract-ocr-ces" "tesseract-ocr-eng" "tesseract-ocr-script-tibt" "jbig2enc")
OPT_LEVEL="3"
JBIG2_LOSSY="1"
REDO_OCR="1"  

echo "Script to process PDF file with Tesseract OCR and compress using jbig2."

MISSING_PKGS=()

for pkg in "${REQUIRED_PKGS[@]}"; do
  if ! dpkg -l | grep -qw "$pkg"; then
    MISSING_PKGS+=("$pkg")
  fi
done

if [ ${#MISSING_PKGS[@]} -ne 0 ]; then
  for missing_pkg in "${MISSING_PKGS[@]}"; do
    if [ "$missing_pkg" == "jbig2enc" ]; then
      echo "Warning: 'jbig2enc' package is missing. It is not available in default Debian/Ubuntu repositories."
      echo "You may add the 'https://notesalexp.org/' repository to your APT sources to install 'jbig2enc'."
      echo "Note: The absence of 'jbig2enc' might affect lossy compression capabilities."
    else
      echo "Missing package: $missing_pkg"
      echo "Attempting to install missing Tesseract OCR packages. Please provide your password if prompted."
      sudo apt-get update
      sudo apt-get install -y "$missing_pkg"
    fi
  done
fi

# Check if a filename is provided
if [ -z "$1" ]; then
  echo "No filename provided. Exiting..."
  exit 1
fi

# Define an array to hold ocrmypdf options
OCR_OPTIONS=(-l "$LANGS" -O"$OPT_LEVEL")
if [ "$JBIG2_LOSSY" -eq "1" ]; then
  OCR_OPTIONS+=("--jbig2-lossy")
fi
if [ "$REDO_OCR" -eq "1" ]; then
  OCR_OPTIONS+=("--redo-ocr")
fi

# Perform OCR using ocrmypdf
if ocrmypdf "${OCR_OPTIONS[@]}" "$1" "${1%.pdf}-OCR.pdf"; then
   echo "OCR completed successfully. Output saved to '${1%.pdf}-OCR.pdf'."
else
  ERROR_CODE=$?
  echo "OCR failed with error code $ERROR_CODE. Please check the input file and try again."
fi