OCR4Linux

OCR4Linux is a versatile text extraction tool that allows you to take a screenshot of a selected area, extract text using OCR, and copy it to the clipboard. It supports both Wayland and X11 sessions and offers multiple language support.

Note: This script is currently only made for Arch Linux. It may work on other arch-based distributions, but it has not been tested yet.

Motivation

I didn't find any easy tool in Linux that does the same thing as the PowerTool app in Windows. This motivated me to create OCR4Linux, a simple and efficient tool to capture screenshots, extract text, and copy it to the clipboard, all in one seamless process.

Features

• Screenshot Capture

  • Wayland support via grimblast
  • X11 support via scrot
  • Configurable screenshot directory

• Text Extraction

  • Interactive language selection via rofi
  • Multi-language OCR support with custom language combinations
  • Automatic language detection fallback
  • Image preprocessing for better accuracy
  • UTF-8 text output

• Clipboard Integration

  • Wayland: wl-copy and cliphist
  • X11: xclip

• Additional Features

  • Interactive language selection menu
  • Optional screenshot retention
  • Comprehensive logging system
  • Command-line interface

Requirements

System Requirements

• Arch Linux or arch-based distribution

• Python 3.x

• yay package manager (will be installed if needed)

• tesseract OCR engine

• tesseract-data-eng English language pack

• tesseract-data-ara Arabic language pack

• If you need any other language other than the above two, search for it using the command:

  sudo pacman -Ss tesseract-data-{lang}

Python Dependencies

• python-pillow
• python-pytesseract

Session-Specific Requirements

• Wayland:
  • grimblast-git
  • wl-clipboard
  • cliphist
  • rofi-wayland
• X11:
  • scrot
  • xclip
  • rofi

Note: rofi is required for the interactive language selection feature.

Installation

1. Clone the repository:

   git clone https://github.com/moheladwy/OCR4Linux.git
   cd OCR4Linux

2. Run the setup script to install the required packages and copy the necessary files to the configuration directory:

   chmod +x setup.sh
   ./setup.sh

Usage

1. Run the main script to take a screenshot, extract text, and copy it to the clipboard:

   chmod +x OCR4Linux.sh
   ./OCR4Linux.sh

2. The script will:

   • With --lang option: Use specified languages directly (bypasses rofi menu)
   • Without --lang option: Display an interactive language selection menu via rofi
   • Allow you to select one or multiple languages for OCR processing
   • Take a screenshot of the selected area after language selection
   • Extract text from the image using the selected languages
   • Copy the extracted text to the clipboard### Language Selection

You have two options for language selection:

Option 1: Command Line (Direct)

Specify languages directly using the --lang option:

• --lang all - Use all available languages
• --lang eng - Use English only
• --lang eng+ara+fra - Use multiple specific languages

Option 2: Interactive Menu (Rofi)

When you run the script without --lang, a rofi menu will appear with:

• ALL: Select all available languages
• Individual languages: Choose specific languages (e.g., eng, ara, fra, deu)
• Multi-select: Hold Ctrl and click to select multiple languages

The selected languages will be used by Tesseract for more accurate text recognition in multi-language documents.

Workflow

The complete OCR4Linux workflow:

1. Language Selection:
   • Command-line specified languages (with --lang) OR
   • Interactive rofi menu displays available languages (without --lang)
2. Language Processing: Selected languages are validated and formatted
3. Screenshot Capture: Area selection and image capture
4. OCR Processing: Text extraction using selected languages
5. Clipboard Integration: Extracted text copied to system clipboard
6. Cleanup: Optional screenshot removal and logging

Command Line Arguments

---

OCR4Linux.sh

Option	Description	Default
-r	Remove screenshot after processing	false
-d DIR	Set screenshot directory	$HOME/Pictures/screenshots
-l	Keep logs	false
--lang LANGUAGES	Specify OCR languages (bypasses rofi)	Interactive selection
-h	Show help message	-

Language Format for --lang:

• Use all for all available languages
• Use + to separate multiple languages (e.g., eng+ara+fra)
• Single languages: eng, ara, fra, etc.

OCR4Linux.py

Option	Description	Required
image_path	Path to input image	Yes
output_path	Path to save extracted text	Yes
--langs <languages>	Specify languages for OCR	No
-l, --list-langs	List available OCR languages	No
-h, --help	Show help message	No

Language Format: Use + to separate multiple languages (e.g., eng+ara+fra)

Examples

---

Using OCR4Linux.sh

# Basic usage (shows interactive rofi menu)
./OCR4Linux.sh

# Direct language specification (bypasses rofi)
./OCR4Linux.sh --lang eng
./OCR4Linux.sh --lang all
./OCR4Linux.sh --lang eng+ara+fra

# Save logs and remove screenshot after processing
./OCR4Linux.sh -l -r

# Custom screenshot directory with logging
./OCR4Linux.sh -d ~/Documents/screenshots -l

# Combine language specification with other options
./OCR4Linux.sh --lang eng -l -r
./OCR4Linux.sh --lang all -d ~/screenshots -l

# Show help
./OCR4Linux.sh -h

Using OCR4Linux.py

# Basic usage (uses all available languages)
python OCR4Linux.py input.png output.txt

# Specify single language
python OCR4Linux.py input.png output.txt --langs eng

# Specify multiple languages
python OCR4Linux.py input.png output.txt --langs eng+ara+fra

# List available languages
python OCR4Linux.py --list-langs

# Show help
python OCR4Linux.py --help

Tips

• Language Selection Options:

  • Command Line: Use --lang for automated/scripted usage

    • --lang all for maximum compatibility
    • --lang eng for English-only documents
    • --lang eng+ara for bilingual documents

  • Interactive Menu: Run without --lang for manual selection

    • Select "ALL" to use all available languages
    • Select specific languages for better performance
    • Use Ctrl+Click to select multiple languages
    • Press Escape to cancel the operation

• Performance Optimization:

  • Use fewer specific languages for faster processing
  • Use --lang all only when document language is unknown
  • Command-line specification is faster than interactive selection

• Keyboard Shortcuts: You can create a keyboard shortcut to run the script for easy access.

  Example for Hyprland users:

  • put the following lines in your hyprland.conf file:

    $OCR4Linux = ~/.config/OCR4Linux/OCR4Linux.sh
    $OCR4Linux_ENG = ~/.config/OCR4Linux/OCR4Linux.sh --lang eng
    
    bind = $mainMod SHIFT, E, exec, $OCR4Linux # OCR4Linux with interactive selection
    bind = $mainMod SHIFT, T, exec, $OCR4Linux_ENG # OCR4Linux with English only
    

  Example for dwm users:

  • put the following lines in your config.h file:

    static const char *ocr4linux[] = { "sh", "-c", "~/.config/OCR4Linux/OCR4Linux.sh", NULL };
    static const char *ocr4linux_eng[] = { "sh", "-c", "~/.config/OCR4Linux/OCR4Linux.sh --lang eng", NULL };
    
    { MODKEY | ShiftMask, XK_e, spawn, {.v = ocr4linux } },      // OCR4Linux interactive
    { MODKEY | ShiftMask, XK_t, spawn, {.v = ocr4linux_eng } },  // OCR4Linux English only

• Language Optimization: For best results:

  • Select only the languages present in your document
  • Use fewer languages for better performance
  • Install additional Tesseract language packs as needed

Files

• OCR4Linux.py: Python script to preprocess the image and extract text using tesseract with support for custom language selection.
• OCR4Linux.sh: Shell script that provides both interactive language selection via rofi and direct command-line language specification, takes a screenshot, passes it to the python script with selected languages, gets the extracted text, and copies it to the clipboard.
• setup.sh: Shell script to install the required packages and copy the necessary files to the configuration directory (run this script the first time you clone the repository only).

Contributing

We welcome contributions from the community to help improve OCR4Linux and make it available for all Linux users and distributions. Whether it's reporting bugs, suggesting new features, or submitting patches, your help is greatly appreciated. Please check out our contributing guidelines to get started.

License

This project is licensed under the MIT License. See the LICENSE file for more details.