OCR Bridge - Tesseract Engine

Tesseract OCR engine implementation for OCR Bridge.

Overview

This package provides a Tesseract OCR engine that integrates with the OCR Bridge architecture. Tesseract is a popular open-source OCR engine developed by Google.

Features

• Multiple Formats: JPEG, PNG, TIFF, PDF
• Multi-page PDFs: Automatic page splitting and merging
• Language Support: 100+ languages via Tesseract language packs
• Configurable: PSM, OEM, and DPI settings
• HOCR Output: Structured XML with bounding boxes

Installation

pip install ocrbridge-tesseract

Note: Tesseract binary must be installed separately:

# Ubuntu/Debian
apt-get install tesseract-ocr tesseract-ocr-eng

# macOS
brew install tesseract

# Windows
# Download from: https://github.com/UB-Mannheim/tesseract/wiki

Usage

The engine is automatically discovered by OCR Bridge via entry points.

Parameters

• lang (str): Language code(s), e.g., "eng", "eng+fra" (default: "eng")
• psm (int): Page segmentation mode 0-13 (default: 3)
• oem (int): OCR engine mode 0-3 (default: 1)
• dpi (int): DPI for PDF conversion, 70-2400 (default: 300)

Example

from pathlib import Path
from ocrbridge.engines.tesseract import TesseractEngine, TesseractParams

engine = TesseractEngine()

# Process with defaults
hocr = engine.process(Path("document.pdf"))

# Process with custom parameters
params = TesseractParams(
    lang="eng+fra",
    psm=6,
    oem=1,
    dpi=300
)
hocr = engine.process(Path("document.pdf"), params)