Compare commits

..
2 Commits
Author SHA1 Message Date
TZGyn 26700c59f4 Make exported stroke thickness match the preview
Strokes were sized in view points but exported in image pixels, so on
retina captures the saved/copied annotation looked several times thinner.
Stroke widths are now converted into image pixels when drawn and scaled
back for the preview, keeping preview and export identical.

💘 Generated with Crush

Assisted-by: Crush:deepseek-v4-flash-0731
2026-09-23 17:28:38 +08:00
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
4 changed files with 43 additions and 22 deletions
+25 -11
View File
@@ -119,33 +119,47 @@ enum LoginItem {
} }
} }
func runOCRSelfTest() { func renderTestImage(_ text: String, fontSize: CGFloat = 56) -> CGImage? {
let text = "Kapture OCR test 1234, hello world" let size = NSSize(width: 1900, height: 220)
let size = NSSize(width: 1600, height: 220)
guard let rep = NSBitmapImageRep( guard let rep = NSBitmapImageRep(
bitmapDataPlanes: nil, pixelsWide: Int(size.width), pixelsHigh: Int(size.height), bitmapDataPlanes: nil, pixelsWide: Int(size.width), pixelsHigh: Int(size.height),
bitsPerSample: 8, samplesPerPixel: 4, hasAlpha: true, isPlanar: false, bitsPerSample: 8, samplesPerPixel: 4, hasAlpha: true, isPlanar: false,
colorSpaceName: .deviceRGB, bytesPerRow: 0, bitsPerPixel: 0 colorSpaceName: .deviceRGB, bytesPerRow: 0, bitsPerPixel: 0
) else { exit(1) } ) else { return nil }
NSGraphicsContext.saveGraphicsState() NSGraphicsContext.saveGraphicsState()
NSGraphicsContext.current = NSGraphicsContext(bitmapImageRep: rep) NSGraphicsContext.current = NSGraphicsContext(bitmapImageRep: rep)
NSColor.white.setFill() NSColor.white.setFill()
NSRect(origin: .zero, size: size).fill() NSRect(origin: .zero, size: size).fill()
let attrs: [NSAttributedString.Key: Any] = [ let attrs: [NSAttributedString.Key: Any] = [
.font: NSFont.boldSystemFont(ofSize: 56), .font: NSFont.systemFont(ofSize: fontSize),
.foregroundColor: NSColor.black, .foregroundColor: NSColor.black,
] ]
text.draw(at: NSPoint(x: 60, y: 75), withAttributes: attrs) text.draw(at: NSPoint(x: 60, y: 75), withAttributes: attrs)
NSGraphicsContext.restoreGraphicsState() NSGraphicsContext.restoreGraphicsState()
return rep.cgImage
}
guard let cgImage = rep.cgImage else { exit(1) } func runOCRSelfTest() {
OCRService.shared.recognize(cgImage) { recognized in let cases: [(expected: String, contains: String)] = [
print("expected: \(text)") ("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)") print("got: \(recognized)")
let ok = recognized == text let ok = recognized.contains(current.contains)
print(ok ? "PASS" : "FAIL") print(ok ? "PASS" : "FAIL")
exit(ok ? 0 : 1) if !ok { exit(1) }
runNext()
} }
}
runNext()
RunLoop.main.run() RunLoop.main.run()
} }
+5
View File
@@ -12,6 +12,11 @@ final class OCRService {
request.recognitionLevel = .accurate request.recognitionLevel = .accurate
request.usesLanguageCorrection = true request.usesLanguageCorrection = true
request.recognitionLanguages = ["en-US", "zh-Hans", "zh-Hant", "ja-JP", "ko-KR"] 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: [:]) let handler = VNImageRequestHandler(cgImage: cgImage, options: [:])
do { do {
+9 -7
View File
@@ -195,14 +195,15 @@ final class CaptureModel: ObservableObject {
return rep.cgImage ?? image return rep.cgImage ?? image
} }
func addStroke(_ points: [CGPoint]) { /// `widthPx` is the stroke thickness in IMAGE PIXELS (the preview
/// converts the slider value so exports match what was drawn).
func addStroke(_ points: [CGPoint], widthPx: CGFloat) {
guard points.count > 1 else { return } guard points.count > 1 else { return }
if tool == .redact { if tool == .redact {
// Redaction: a filled black bar from start to end, as thick as the // Redaction: a filled black bar from start to end.
// current stroke width (drawn as a rect for clean edges). strokes.append(Stroke(points: points, width: max(widthPx, 12), color: .black, isRedaction: true))
strokes.append(Stroke(points: points, width: max(strokeWidth, 12), color: .black, isRedaction: true))
} else { } else {
strokes.append(Stroke(points: points, width: strokeWidth, color: drawColor)) strokes.append(Stroke(points: points, width: widthPx, color: drawColor))
} }
redoStack = [] redoStack = []
} }
@@ -575,7 +576,8 @@ struct ScreenshotPane: View {
for point in stroke.points.dropFirst() { for point in stroke.points.dropFirst() {
path.addLine(to: CGPoint(x: origin.x + point.x * scale, y: origin.y + point.y * scale)) path.addLine(to: CGPoint(x: origin.x + point.x * scale, y: origin.y + point.y * scale))
} }
context.stroke(path, with: .color(Color(nsColor: stroke.color)), lineWidth: stroke.width) // stroke.width is in image pixels; scale to view points
context.stroke(path, with: .color(Color(nsColor: stroke.color)), lineWidth: stroke.width * scale)
} }
} }
} }
@@ -636,7 +638,7 @@ struct ScreenshotPane: View {
.onEnded { _ in .onEnded { _ in
if model.tool != .select { if model.tool != .select {
if currentStroke.count > 1 { if currentStroke.count > 1 {
model.addStroke(currentStroke) model.addStroke(currentStroke, widthPx: model.strokeWidth / max(scale, 0.01))
} }
currentStroke = [] currentStroke = []
} else { } else {
+2 -2
View File
@@ -108,9 +108,9 @@ cat > "$APP/Contents/Info.plist" <<'PLIST'
<key>CFBundleIconFile</key> <key>CFBundleIconFile</key>
<string>Icon</string> <string>Icon</string>
<key>CFBundleShortVersionString</key> <key>CFBundleShortVersionString</key>
<string>1.9.0</string> <string>1.9.2</string>
<key>CFBundleVersion</key> <key>CFBundleVersion</key>
<string>90</string> <string>92</string>
<key>LSUIElement</key> <key>LSUIElement</key>
<true/> <true/>
<key>NSHighResolutionCapable</key> <key>NSHighResolutionCapable</key>