TZGyn/kapture
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 52 Wiki Activity

Retry OCR on upscaled images when nothing is read

Browse Source 
Tiny UI text (around 7pt) reads as nothing at native size — verified
locally. When a pass returns fewer than three characters, OCR retries at
2x and 3x and keeps the fullest result. The smoke test covers a tiny-text
case so the fallback stays working.

💘 Generated with Crush

Assisted-by: Crush:deepseek-v4-flash-0731
This commit is contained in:
TZGyn
2026-09-30 11:44:26 +08:00
parent 2e971d5f8d
commit 04b84e188a
3 changed files with 73 additions and 36 deletions
Download Patch File Download Diff File Expand all files Collapse all files
Sources/Kapture/App.swift
+13 -8
View File
@@ -123,8 +123,8 @@ enum LoginItem {
}
}
func renderTestImage(_ text: String, fontSize: CGFloat = 56) -> CGImage? {
let size = NSSize(width: 1900, height: 220)
func renderTestImage(_ text: String, fontSize: CGFloat = 56, width: CGFloat = 1900, height: CGFloat = 220) -> CGImage? {
let size = NSSize(width: width, height: height)
guard let rep = NSBitmapImageRep(
bitmapDataPlanes: nil, pixelsWide: Int(size.width), pixelsHigh: Int(size.height),
bitsPerSample: 8, samplesPerPixel: 4, hasAlpha: true, isPlanar: false,
@@ -138,15 +138,20 @@ func renderTestImage(_ text: String, fontSize: CGFloat = 56) -> CGImage? {
.font: NSFont.systemFont(ofSize: fontSize),
.foregroundColor: NSColor.black,
]
text.draw(at: NSPoint(x: 60, y: 75), withAttributes: attrs)
// Centre the text vertically so tiny canvases still contain it.
let textHeight = ("Ag" as NSString).size(withAttributes: attrs).height
let y = max(4, (size.height - textHeight) / 2)
text.draw(at: NSPoint(x: 20, y: y), withAttributes: attrs)
NSGraphicsContext.restoreGraphicsState()
return rep.cgImage
}
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"),
let cases: [(expected: String, contains: String, fontSize: CGFloat, width: CGFloat, height: CGFloat)] = [
("Kapture OCR test 1234, hello world", "Kapture OCR test 1234", 56, 1900, 220),
("你好世界，这是一个测试", "你好世界", 56, 1900, 220),
// Tiny text: needs the upscale retry to read anything at all.
("会话导出并发冲突报错 export fail", "导出", 7, 320, 40),
]
var index = 0
@@ -154,9 +159,9 @@ func runOCRSelfTest() {
guard index < cases.count else { exit(0) }
let current = cases[index]
index += 1
guard let image = renderTestImage(current.expected) else { exit(1) }
guard let image = renderTestImage(current.expected, fontSize: current.fontSize, width: current.width, height: current.height) else { exit(1) }
OCRService.shared.recognize(image) { recognized in
print("case: \(current.expected)")
print("case: \(current.expected) @\(Int(current.fontSize))pt")
print("got: \(recognized)")
let ok = recognized.contains(current.contains)
print(ok ? "PASS" : "FAIL")