VisionOCR.swift

import Foundation
import Vision
import AppKit

func performOCR(on image: NSImage, completion: @escaping ([String: Any]) -> Void) {
    guard let cgImage = image.cgImage(forProposedRect: nil, context: nil, hints: nil) else {
        print("Error: Unable to get CGImage from NSImage.")
        exit(1)
    }

    let requestHandler = VNImageRequestHandler(cgImage: cgImage, options: [:])
    let request = VNRecognizeTextRequest { (request, error) in
        if let error = error {
            print("Error: \(error.localizedDescription)")
            exit(1)
        }

        guard let observations = request.results as? [VNRecognizedTextObservation] else {
            print("Error: Unable to get recognized text.")
            exit(1)
        }

        var jsonArray: [[String: Any]] = []

        for observation in observations {
            if let topCandidate = observation.topCandidates(1).first {
                let text = topCandidate.string
                let boundingBox = observation.boundingBox

                let position = [
                    "x": boundingBox.origin.x,
                    "y": boundingBox.origin.y,
                    "width": boundingBox.size.width,
                    "height": boundingBox.size.height
                ]

                jsonArray.append([
                    "text": text,
                    "position": position
                ])
            }
        }

        completion(["result": jsonArray])
    }

    request.recognitionLevel = .accurate
    request.usesLanguageCorrection = true

    try? requestHandler.perform([request])
}

func loadImage(at path: String) -> NSImage? {
    let url = URL(fileURLWithPath: path)
    return NSImage(contentsOf: url)
}

guard CommandLine.argc == 2 else {
    print("Usage: \(CommandLine.arguments[0]) path/to/image")
    exit(1)
}

let imagePath = CommandLine.arguments[1]

guard let image = loadImage(at: imagePath) else {
    print("Error: Unable to load image at path: \(imagePath)")
    exit(1)
}

performOCR(on: image) { json in
    let jsonData = try! JSONSerialization.data(withJSONObject: json, options: .prettyPrinted)
    let jsonString = String(data: jsonData, encoding: .utf8)!
    print(jsonString)
    exit(0)
}

RunLoop.main.run()