Compare commits

..
1 Commits
Author SHA1 Message Date
TZGyn 195c6a9b1f Enable automatic language detection so CJK text is recognized
Vision defaulted to the first listed language and turned Chinese into
Latin garbage (你好世界 became 177WR). automaticallyDetectsLanguage fixes
it; the smoke test now covers Chinese so it cannot regress.

💘 Generated with Crush

Assisted-by: Crush:deepseek-v4-flash-0731
2026-09-22 10:54:54 +08:00
3 changed files with 34 additions and 15 deletions
+27 -13
View File
@@ -119,33 +119,47 @@ enum LoginItem {
}
}
func runOCRSelfTest() {
let text = "Kapture OCR test 1234, hello world"
let size = NSSize(width: 1600, height: 220)
func renderTestImage(_ text: String, fontSize: CGFloat = 56) -> CGImage? {
let size = NSSize(width: 1900, height: 220)
guard let rep = NSBitmapImageRep(
bitmapDataPlanes: nil, pixelsWide: Int(size.width), pixelsHigh: Int(size.height),
bitsPerSample: 8, samplesPerPixel: 4, hasAlpha: true, isPlanar: false,
colorSpaceName: .deviceRGB, bytesPerRow: 0, bitsPerPixel: 0
) else { exit(1) }
) else { return nil }
NSGraphicsContext.saveGraphicsState()
NSGraphicsContext.current = NSGraphicsContext(bitmapImageRep: rep)
NSColor.white.setFill()
NSRect(origin: .zero, size: size).fill()
let attrs: [NSAttributedString.Key: Any] = [
.font: NSFont.boldSystemFont(ofSize: 56),
.font: NSFont.systemFont(ofSize: fontSize),
.foregroundColor: NSColor.black,
]
text.draw(at: NSPoint(x: 60, y: 75), withAttributes: attrs)
NSGraphicsContext.restoreGraphicsState()
return rep.cgImage
}
guard let cgImage = rep.cgImage else { exit(1) }
OCRService.shared.recognize(cgImage) { recognized in
print("expected: \(text)")
print("got: \(recognized)")
let ok = recognized == text
print(ok ? "PASS" : "FAIL")
exit(ok ? 0 : 1)
func runOCRSelfTest() {
let cases: [(expected: String, contains: String)] = [
("Kapture OCR test 1234, hello world", "Kapture OCR test 1234"),
("你好世界,这是一个测试", "你好世界"),
]
var index = 0
func runNext() {
guard index < cases.count else { exit(0) }
let current = cases[index]
index += 1
guard let image = renderTestImage(current.expected) else { exit(1) }
OCRService.shared.recognize(image) { recognized in
print("case: \(current.expected)")
print("got: \(recognized)")
let ok = recognized.contains(current.contains)
print(ok ? "PASS" : "FAIL")
if !ok { exit(1) }
runNext()
}
}
runNext()
RunLoop.main.run()
}
+5
View File
@@ -12,6 +12,11 @@ final class OCRService {
request.recognitionLevel = .accurate
request.usesLanguageCorrection = true
request.recognitionLanguages = ["en-US", "zh-Hans", "zh-Hant", "ja-JP", "ko-KR"]
// Without this, Vision defaults to the first listed language and
// renders CJK as Latin garbage (verified: 你好世界 -> "177WR").
if #available(macOS 13.0, *) {
request.automaticallyDetectsLanguage = true
}
let handler = VNImageRequestHandler(cgImage: cgImage, options: [:])
do {
+2 -2
View File
@@ -108,9 +108,9 @@ cat > "$APP/Contents/Info.plist" <<'PLIST'
<key>CFBundleIconFile</key>
<string>Icon</string>
<key>CFBundleShortVersionString</key>
<string>1.9.0</string>
<string>1.9.1</string>
<key>CFBundleVersion</key>
<string>90</string>
<string>91</string>
<key>LSUIElement</key>
<true/>
<key>NSHighResolutionCapable</key>