Add option to disable l to I correction

Signed-off-by: Ethan Dye <[email protected]>

Sources/macSubtitleOCR/Subtitles/SubtitleProcessor.swift

@@ -52,21 +52,23 @@ struct SubtitleProcessor {
     private let language: String
     private let fastMode: Bool
     private let disableLanguageCorrection: Bool
+    private let disableICorrection: Bool
     private let forceOldAPI: Bool
     private let outputDirectory: String
     private let maxConcurrentTasks: Int
     private let logger = Logger(subsystem: "github.ecdye.macSubtitleOCR", category: "SubtitleProcessor")
 
     init(subtitles: [Subtitle], trackNumber: Int, invert: Bool, saveImages: Bool, language: String, fastMode: Bool,
-         disableLanguageCorrection: Bool,
-         forceOldAPI: Bool, outputDirectory: String, maxConcurrentTasks: Int) {
+         disableLanguageCorrection: Bool, disableICorrection: Bool, forceOldAPI: Bool, outputDirectory: String,
+         maxConcurrentTasks: Int) {
         self.subtitles = subtitles
         self.trackNumber = trackNumber
         self.invert = invert
         self.saveImages = saveImages
         self.language = language
         self.fastMode = fastMode
         self.disableLanguageCorrection = disableLanguageCorrection
+        self.disableICorrection = disableICorrection
         self.forceOldAPI = forceOldAPI
         self.outputDirectory = outputDirectory
         self.maxConcurrentTasks = maxConcurrentTasks
@@ -103,8 +105,12 @@ struct SubtitleProcessor {
                     }
 
                     let (subtitleText, subtitleLines) = await recognizeText(from: subImage)
-                    let pattern = #"\bl\b"# // Replace l with I when it's a single character
-                    subtitle.text = subtitleText.replacingOccurrences(of: pattern, with: "I", options: .regularExpression)
+                    if language.contains("en"), !disableICorrection {
+                        let pattern = #"\bl\b"# // Replace l with I when it's a single character
+                        subtitle.text = subtitleText.replacingOccurrences(of: pattern, with: "I", options: .regularExpression)
+                    } else {
+                        subtitle.text = subtitleText
+                    }
                     subtitle.imageData = nil // Clear the image data to save memory
 
                     let jsonOut = SubtitleJSONResult(index: subIndex, lines: subtitleLines, text: subtitleText)

Sources/macSubtitleOCR/macSubtitleOCR.swift

@@ -26,14 +26,13 @@ struct ExperimentalOptions: ParsableArguments {
     var disableLanguageCorrection = false
 }
 
-// The main struct representing the macSubtitleOCR command-line tool.
 @main
 struct macSubtitleOCR: AsyncParsableCommand {
     // MARK: - Properties
 
     static let configuration = CommandConfiguration(
         commandName: "macSubtitleOCR",
-        abstract: "macSubtitleOCR - Convert bitmap subtitles into SubRip format using the macOS OCR engine")
+        abstract: "macSubtitleOCR - Convert bitmap subtitles into SubRip format using the macOS Vision framework")
 
     @Argument(help: "Input subtitle file (supported formats: .sup, .sub, .idx, .mkv)")
     var input: String
@@ -63,6 +62,9 @@ struct macSubtitleOCR: AsyncParsableCommand {
     @Flag(help: "Use FFmpeg decoder")
     var ffmpegDecoder = false
 
+    @Flag(help: "Disable correction of 'l' to 'I' in OCR results")
+    var disableICorrection = false
+
     @OptionGroup(title: "Experimental Options", visibility: .hidden)
     var experimentalOptions: ExperimentalOptions
 
@@ -164,6 +166,7 @@ struct macSubtitleOCR: AsyncParsableCommand {
             language: languages,
             fastMode: experimentalOptions.fastMode,
             disableLanguageCorrection: experimentalOptions.disableLanguageCorrection,
+            disableICorrection: disableICorrection,
             forceOldAPI: experimentalOptions.forceOldAPI,
             outputDirectory: outputDirectory,
             maxConcurrentTasks: maxThreads)