ocr-stringdist

OCR-StringDist

A Python library for string distance calculations that account for common OCR (optical character recognition) errors.

Documentation: https://niklasvonm.github.io/ocr-stringdist/

Overview

OCR-StringDist provides specialized string distance algorithms that accommodate for optical character recognition (OCR) errors. Unlike traditional string comparison algorithms, OCR-StringDist considers common OCR confusions (like "0" vs "O", "6" vs "G", etc.) when calculating distances between strings.

Note: This project is in early development. APIs may change in future releases.

Installation

pip install ocr-stringdist

Features

• Weighted Levenshtein Distance: An adaptation of the classic Levenshtein algorithm with custom substitution costs for character pairs that are commonly confused in OCR models, including efficient batch processing.
• Unicode Support: Arbitrary unicode strings can be compared.
• Substitution of Multiple Characters: Not just character pairs, but string pairs may be substituted, for example the Korean syllable "이" for the two letters "OI".
• Pre-defined OCR Distance Map: A built-in distance map for common OCR confusions (e.g., "0" vs "O", "1" vs "l", "5" vs "S").
• Best Match Finder: Utility function find_best_candidate to efficiently find the best matching string from a collection of candidates using any specified distance function (including the library's OCR-aware ones).

Usage

Weighted Levenshtein Distance

import ocr_stringdist as osd

# Using default OCR distance map
distance = osd.weighted_levenshtein_distance("OCR5", "OCRS")
print(f"Distance between 'OCR5' and 'OCRS': {distance}")  # Will be less than 1.0

# Custom cost map
custom_map = {("In", "h"): 0.5}
distance = osd.weighted_levenshtein_distance(
    "hi", "Ini",
    cost_map=custom_map,
    symmetric=True,
)
print(f"Distance with custom map: {distance}")

Finding the Best Candidate

import ocr_stringdist as osd

s = "apple"
candidates = ["apply", "apples", "orange", "appIe"]  # 'appIe' has an OCR-like error

def ocr_aware_distance(s1: str, s2: str) -> float:
    return osd.weighted_levenshtein_distance(s1, s2, cost_map={("l", "I"): 0.1})

best_candidate, best_dist = osd.find_best_candidate(s, candidates, ocr_aware_distance)
print(f"Best candidate for '{s}' is '{best_candidate}' with distance {best_dist}")
# Output: Best candidate for 'apple' is 'appIe' with distance 0.1

Acknowledgements

This project is inspired by jellyfish, providing the base implementations of the algorithms used here.